Quick & Indepth

C
With Data Structures

Sudripta Nandy

1
Quick & Indepth C With Data Structures

ABOUT THE AUTHOR

Sudripta has almost 20 years of experience working on various computer technologies ranging from
desktop computers to Mainframes and to data centers. He has worked on multiple projects for many
Fortune 500 companies and in areas like operating system, media technologies, content protection,
data processing, backup technologies, services, setup and application development. Professionally, he
has spent his spare time educating and training upcoming young minds to the intricacies of
programming. In the personal front, he divides his time playing with his sports fanatic son, being a
couch potato and serving 'Her Majesty' - His wife :-). He can be reached at [email protected].

To receive a soft copy of the program source codes discussed in this book, you may email a request to
the author with the subject line as 'Quick and Indepth C - Source codes'. Please provide your name,
your country of residence, the dealer details and the purchase date in the body of the email.

2
 ACKNOWLEDGMENTS
I may be the author of this book, but, it would have been impossible without the love and support of
many people around me. People who helped me grow up, get educated, guided me, inspired me, made
this book a possibility. There are so many people who touched my life, that I would need hundreds of
pages to dedicate them individually. If someone is left out over here, do remember that you reside in the
most precious place within me – my heart.

The first person I want to dedicate this book to, is my late grandfather, whose life principles I carry to
this day. I can relate many of my characteristics to him – some may not suit this modern era but, I am
proud of them. The next person who helped me shape up as a person and still doing so, is my mother.
She spent so many sleepless nights when I got those asthma attacks. I slept because she stayed awake,
I am alive because she stayed strong. She is my pillar of strength.

When I think about my life, its incomplete without my father – the person who never got angry on me,
who died working so that I could live. The most simple yet straight forward person I know. He left a void
in my heart which can never be filled.

The person who fills my life with the warmth I look for, is my son. Being a father is the best feeling a
person can have. He has helped me grow so much as a person and as a father. The next person who
completes me is my wife – the best friend in my entire life. The love and support, I get from her help me
in life’s every journey I take.

A person’s life is incomplete without a sister, and I am lucky to have two. The biggest support pillars of
my life, the persons I can always count on when in need. My grandmother is the person in whose lap I
started off my journey called life. My niece is the person who gave me the first feelings of being the
older generation – and what a beautiful feeling. Still visualize those chubby cheeks and big round eyes.

There are many others who played an important role in my life and guided me towards this journey. My
father-in-law, mother-in-law, moni, mesho, mezdabhai, brothers-in-law, sister-in-law, nephews, Sumit –
all helped me become the person I am today. They may not have directly helped me in writing this book
but, have helped the author within me evolve.

This book is also dedicated to the many friends and colleagues I had over the years. I learned from
every colleague (special mention A Chakrabarti, A Ray) I had during my entire professional career. The
work I did within my organization also helped me grow as a professional and come to this point where I
can confidently write a book.

3
© 2018 Sudripta Nandy, Salt Lake, Kolkata-700091, India, All Rights Reserved.

No part of this book, including texts, source codes, drawings, images, interior design, cover design,
icons, may be reproduced or transmitted in any form except with the permission of the author 'Sudripta
Nandy' of Salt Lake, Kolkata-700091, India.

Legal Disclaimer: The author of this book has made every attempt to ensure the accuracy and reliability
of the information provided in this book. However, the information is provided 'as is' without warranty of
any kind. The book provides reference information on the subject discussed, and the use and
application of the provided information is entirely dependent on the discretion of the reader. The author
and the publisher assume no responsibility (and liability) for the use of such information in whatever
possible manner. The author and the publisher will not be responsible for any loss or damage of
whatever nature (direct, indirect, consequential or other) whether arising in contract, tort or otherwise,
which may arise as a result of the use of (or inability to use) this book, or from the use of (or failure to
use) the information in this book.

The author and publisher also assume no responsibility regarding the web links and their contents
specified in this book. The web links are provided to help the reader have easy and quick availability of
matter discussed in this book. The web links may further redirect the user to other web locations. The
web links and their locations, the redirected locations and their contents are provided ‘as is’ without
warranty of any kind. The author and publisher of this book hold no responsibility (and liability) for the
web links and their locations, their redirected locations and their contents. Their access, use, view,
download, execution are entirely at the sole discretion of the person doing so. The author and the
publisher will not be responsible for any loss or damage of whatever nature (direct, indirect,
consequential or other) whether arising in contract, tort or otherwise, which may arise as a result of the
use of (or inability to use) those links and locations, or from the use of (or failure to use) the contents
available through those links or locations.

4
 CONTENTS
INTRODUCTION TO C..................................................................................................13
1.1 IMPORTANCE OF C............................................................................................................................13
1.2 WRITING PROGRAMS........................................................................................................................13
1.3 OUR FIRST PROGRAM.......................................................................................................................14
1.4 STRUCTURE OF C PROGRAMS..........................................................................................................16
1.5 EXECUTING A 'C' PROGRAM.............................................................................................................16
1.6 IDENTIFIERS IN C.............................................................................................................................17
DATA TYPES, VARIABLES AND CONSTANTS................................................................18
2.1 INTRODUCTION.................................................................................................................................18
2.2 CHARACTER REPRESENTATION........................................................................................................18
2.3 TOKENS IN C......................................................................................................................................19
2.4 KEYWORDS.........................................................................................................................................19
2.5 CONSTANTS.......................................................................................................................................20
2.5.1 INTEGER CONSTANTS...............................................................................................................20
2.5.2 REAL CONSTANTS.....................................................................................................................21
2.5.3 SINGLE CHARACTER CONSTANTS............................................................................................21
2.5.4 STRING CONSTANTS.................................................................................................................22
2.6 IDENTIFIERS......................................................................................................................................22
2.7 VARIABLES.........................................................................................................................................22
2.7.1 VARIABLE DECLARATION..........................................................................................................23
2.8 DATA TYPES........................................................................................................................................23
2.8.1 PRIMARY DATA TYPES...............................................................................................................23
2.8.2 SIZE AND RANGE OF PRIMARY DATA TYPES............................................................................25
2.8.3 DERIVED DATA TYPES................................................................................................................26
2.8.4 ALIAS DATA TYPES....................................................................................................................26
2.8.5 void DATA TYPE..........................................................................................................................28
2.9 STORAGE CLASSES............................................................................................................................28
2.10 DEFINING SYMBOLIC CONSTANTS AND MACROS.........................................................................29
2.11 DECLARING A VARIABLE AS A CONSTANT....................................................................................31
2.12 VOLATILE VARIABLES......................................................................................................................31

5
Quick & Indepth C With Data Structures

2.13 DATA OVERFLOW AND UNDERFLOW................................................................................................32

2.14 SAMPLE PROGRAM..........................................................................................................................32
OPERATORS AND OPERATIONS....................................................................................34
3.1 INTRODUCTION.................................................................................................................................34
3.2 ARITHMETIC OPERATORS.................................................................................................................34
3.3 INCREMENT AND DECREMENT OPERATORS.....................................................................................35
3.4 ASSIGNMENT OPERATORS................................................................................................................37
3.5 BITWISE OPERATORS........................................................................................................................37
3.6 RELATIONAL OPERATORS..................................................................................................................40
3.7 LOGICAL OPERATORS........................................................................................................................41
3.8 CONDITIONAL (TERNARY) OPERATOR..............................................................................................42
3.9 SPECIAL OPERATORS........................................................................................................................43
3.9.1 COMMA OPERATOR....................................................................................................................43
3.9.2 sizeof OPERATOR.......................................................................................................................43
3.10 ARITHMETIC EXPRESSIONS...........................................................................................................43
3.11 TYPE CONVERSIONS.......................................................................................................................45
3.12 OPERATOR PRECEDENCE AND ASSOCIATIVITY.............................................................................47
3.13 SAMPLE PROGRAMS........................................................................................................................48
INPUT, OUTPUT AND FORMAT SPECIFIERS................................................................52
4.1 INTRODUCTION.................................................................................................................................52
4.2 READING INPUT.................................................................................................................................52
4.2.1 SIMPLE INPUT FUNCTIONS......................................................................................................52
4.2.2 FORMATTED INPUT...................................................................................................................53
4.3 FORMAT SPECIFIERS.........................................................................................................................54
4.3.1 FORMATTED INPUT [EXAMPLES]..............................................................................................56
4.4 GENERATING OUTPUT.......................................................................................................................56
4.4.1 SIMPLE OUTPUT FUNCTIONS..................................................................................................56
4.4.2 FORMATTED OUTPUT................................................................................................................57
4.5 SAMPLE PROGRAMS..........................................................................................................................58
ARRAYS........................................................................................................................61
5.1 INTRODUCTION.................................................................................................................................61
5.2 ONE-DIMENSIONAL ARRAYS.............................................................................................................62

6
 5.2.1 INITIALIZATION OF ONE-DIMENSIONAL ARRAYS...................................................................62
5.2.2 STRINGS....................................................................................................................................63
5.3 TWO-DIMENSIONAL ARRAYS............................................................................................................64
5.3.1 INITIALIZATION OF TWO-DIMENSIONAL ARRAYS...................................................................68
5.3.2 STRING ARRAYS........................................................................................................................69
5.4 MULTI-DIMENSIONAL ARRAYS..........................................................................................................69
5.5 SAMPLE PROGRAMS..........................................................................................................................70
CONDITIONAL STATEMENTS AND BRANCHING..........................................................76
6.1 INTRODUCTION.................................................................................................................................76
6.2 DECISION CONTROL USING if STATEMENT......................................................................................76
6.3 DECISION CONTROL USING if-else STATEMENT..............................................................................78
6.4 DECISION CONTROL USING if-else LADDER STATEMENTS.............................................................80
6.5 DECISION CONTROL USING NESTED if STATEMENTS.....................................................................82
6.6 DECISION CONTROL USING switch STATEMENT.............................................................................84
6.7 DECISION CONTROL USING CONDITIONAL (TERNARY) OPERATOR...............................................87
6.8 DECISION CONTROL USING goto STATEMENT................................................................................88
6.9 SAMPLE PROGRAMS..........................................................................................................................89
CONDITIONAL STATEMENTS WITH LOOPS..................................................................94
7.1 INTRODUCTION.................................................................................................................................94
7.2 THE while LOOP..................................................................................................................................95
7.3 THE do while LOOP............................................................................................................................96
7.4 THE for LOOP.....................................................................................................................................97
7.5 NESTED LOOPS................................................................................................................................100
7.6 JUMPING WITHIN LOOPS................................................................................................................102
7.6.1 THE break STATEMENT...........................................................................................................102
7.6.2 THE continue STATEMENT......................................................................................................105
7.7 CONCISE TEST CONDITIONS.........................................................................................................108
7.8 SAMPLE PROGRAMS........................................................................................................................108
POINTERS..................................................................................................................114
8.1 INTRODUCTION...............................................................................................................................114
8.2 DECLARING, INITIALIZING AND ASSIGNING POINTERS..............................................................116
8.3 ACCESSING A VARIABLE USING ITS POINTER..............................................................................118

7
Quick & Indepth C With Data Structures

8.4 ARRAY OF POINTERS AND POINTER TO AN ARRAY.......................................................................119

8.4.1 ARRAY OF POINTERS..............................................................................................................119
8.4.2 POINTER TO AN ARRAY...........................................................................................................120
8.5 POINTER ARITHMETIC....................................................................................................................122
8.6 POINTER TO ANOTHER POINTER....................................................................................................129
8.7 SAMPLE PROGRAMS........................................................................................................................131
CHARACTER STRING OPERATIONS...........................................................................136
9.1 INTRODUCTION...............................................................................................................................136
9.2 UNICODE CHARACTERS..................................................................................................................138
9.3 ACCEPTING STRINGS FROM USER.................................................................................................139
9.4 PRINTING STRINGS TO CONSOLE..................................................................................................142
9.5 ARITHMETIC OPERATIONS WITH CHARACTERS...........................................................................143
9.6 ARITHMETIC OPERATIONS WITH STRINGS...................................................................................144
9.7 STRING PROCESSING FUNCTIONS................................................................................................144
9.7.1 GENERAL STRING PROCESSING FUNCTIONS.......................................................................144
9.7.2 MEMORY COPY AND COMPARISON FUNCTIONS...................................................................150
9.7.3 SINGLE CHARACTER PROCESSING FUNCTIONS..................................................................151
9.7.4 STRING CONVERSION FUNCTIONS.......................................................................................153
9.8 SAMPLE PROGRAMS........................................................................................................................158
STRUCTURES AND UNIONS.......................................................................................162
10.1 INTRODUCTION............................................................................................................................162
10.2 DEFINING A STRUCTURE..............................................................................................................162
10.3 ACCESSING MEMBERS.................................................................................................................164
10.4 STRUCTURE INITIALIZATION......................................................................................................166
10.5 ARRAY OF STRUCTURE.................................................................................................................167
10.6 POINTER TO STRUCTURE.............................................................................................................169
10.7 NESTED STRUCTURES..................................................................................................................170
10.8 BIT FIELDS....................................................................................................................................173
10.9 UNIONS.........................................................................................................................................176
USER DEFINED FUNCTIONS......................................................................................179
11.1 INTRODUCTION............................................................................................................................179
11.2 THE STRUCTURE OF A C FUNCTION............................................................................................180

8
 11.2.1 FUNCTION NAME..................................................................................................................180
11.2.2 FUNCTION RETURN TYPE AND VALUES..............................................................................181
11.2.3 FUNCTION ARGUMENTS......................................................................................................182
11.2.4 FUNCTION BODY...................................................................................................................183
11.2.5 EXAMPLE PROGRAMS...........................................................................................................183
11.3 FUNCTION VARIABLES: SCOPE AND LIFETIME...........................................................................184
11.3.1 AUTOMATIC VARIABLES.......................................................................................................185
11.3.2 GLOBAL AND EXTERNAL VARIABLES...................................................................................186
11.3.3 STATIC VARIABLES...............................................................................................................189
11.3.4 REGISTER VARIABLES..........................................................................................................191
11.4 FUNCTIONS WITH POINTER ARGUMENTS..................................................................................191
11.4.1 FUNCTION POINTER AS ARGUMENT...................................................................................193
11.5 FUNCTIONS WITH ARRAY ARGUMENTS......................................................................................195
11.6 FORWARD DECLARATION OF FUNCTIONS...................................................................................197
11.7 VARIABLE NUMBER OF ARGUMENTS...........................................................................................199
11.8 RECURSION...................................................................................................................................200
11.9 SAMPLE PROGRAMS......................................................................................................................203
FILE MANAGEMENT...................................................................................................210
12.1 INTRODUCTION.............................................................................................................................210
12.2 CREATING AND OPENING FILE.....................................................................................................211
12.3 CLOSING A FILE............................................................................................................................212
12.4 WRITING-TO AND READING-FROM A TEXTUAL FILE..................................................................213
12.4.1 WRITING AND READING SINGLE CHARACTERS.................................................................213
12.4.2 WRITING AND READING INTEGERS.....................................................................................215
12.4.3 WRITING AND READING STRINGS.......................................................................................215
12.4.4 WRITING AND READING FORMATTED DATA........................................................................216
12.5 WRITING-TO AND READING-FROM A BINARY FILE.....................................................................221
12.6 MISCELLANEOUS FILE MANAGEMENT FUNCTIONS....................................................................226
DYNAMIC MEMORY ALLOCATION..............................................................................228
13.1 INTRODUCTION.............................................................................................................................228
13.2 C PROGRAM – MEMORY LAYOUT...................................................................................................228
13.3 ALLOCATING MEMORY..................................................................................................................230
13.4 ALTERING ALLOCATED MEMORY SIZE..........................................................................................232

9
Quick & Indepth C With Data Structures

13.5 RELEASING ALLOCATED SPACE....................................................................................................232

13.6 SAMPLE PROGRAM........................................................................................................................233
13.7 THINGS TO REMEMBER.................................................................................................................234
COMMAND LINE ARGUMENTS...................................................................................236
14.1 INTRODUCTION.............................................................................................................................236
14.2 ‘main’ FUNCTION REVISITED.......................................................................................................236
14.3 SAMPLE PROGRAMS......................................................................................................................237
PREPROCESSOR DIRECTIVES....................................................................................241
15.1 INTRODUCTION.............................................................................................................................241
15.2 FILE INCLUSION DIRECTIVES......................................................................................................241
15.3 MACRO DIRECTIVES......................................................................................................................242
15.3.1 SIMPLE MACRO SUBSTITUTION..........................................................................................242
15.3.2 MACRO SUBSTITUTION USING ARGUMENTS.....................................................................244
15.3.3 NESTED MACRO SUBSTITUTION.........................................................................................246
15.3.4 STANDARD PREDEFINED MACROS.......................................................................................247
15.3.5 UNDEFINING A MACRO.........................................................................................................248
15.4 COMPILER CONTROL DIRECTIVES...............................................................................................248
15.4.1 CONDITIONAL PREPROCESSOR STATEMENTS....................................................................248
15.4.2 Pragma DIRECTIVES............................................................................................................251
15.4.3 DIAGNOSTIC PREPROCESSOR STATEMENTS......................................................................253
15.5 SAMPLE PROGRAM........................................................................................................................253
DEPRECATED FUNCTIONS.........................................................................................257

BEST PROGRAMMING PRACTICES............................................................................259

17.1 INTRODUCTION.............................................................................................................................259
17.1.1 DOCUMENTATION.................................................................................................................259
17.1.2 AVOID USING DEPRECATED FUNCTIONS............................................................................259
17.1.3 TABS......................................................................................................................................259
17.1.4 INDENTATION.......................................................................................................................259
17.1.5 LINE LENGTH........................................................................................................................260
17.1.6 BLANK LINES........................................................................................................................260
17.1.7 FILE HEADER.........................................................................................................................260
17.1.8 FUNCTION HEADER..............................................................................................................260

10
 17.1.9 BRACES.................................................................................................................................260
17.1.10 PREVENT MULTIPLE INCLUSIONS.....................................................................................261
17.1.11 FUNCTION POINTERS........................................................................................................261
17.1.12 USE OF PARENTHESIS.......................................................................................................261
17.1.13 BOOLEAN COMPARISONS..................................................................................................261
17.1.14 COMPARING CONSTANTS..................................................................................................262
17.1.15 ALWAYS CHECK RETURN VALUES......................................................................................262
17.1.16 BLOCK STATEMENTS..........................................................................................................262
17.1.17 SWITCH STATEMENTS........................................................................................................262
17.1.18 DISABLING OUT CODE.......................................................................................................263
17.1.19 SHORT FUNCTIONS ARE BEAUTIFUL................................................................................263
17.1.20 LIMIT EXIT POINTS FROM FUNCTIONS............................................................................263
17.1.21 INITIALIZE VARIABLES......................................................................................................265
17.1.22 VARIABLE NAMING CONVENTIONS...................................................................................265
LINKED LISTS............................................................................................................266
18.1 INTRODUCTION.............................................................................................................................266
18.2 LINKED LIST TYPES.......................................................................................................................267
18.2.1 SINGLE LINEAR LINKED LIST...............................................................................................267
18.2.2 DOUBLE LINEAR LINKED LIST..............................................................................................272
18.2.3 SINGLE CIRCULAR LINKED LIST..........................................................................................277
18.2.4 DOUBLE CIRCULAR LINKED LIST.........................................................................................281
STACKS AND QUEUES................................................................................................285
19.1 INTRODUCTION.............................................................................................................................285
19.2 STACKS..........................................................................................................................................285
19.2.1 STACK USING ARRAY............................................................................................................286
19.2.2 STACK USING LINKED LIST..................................................................................................287
19.3 QUEUES..........................................................................................................................................289
19.3.1 QUEUE USING ARRAY...........................................................................................................289
19.3.2 QUEUE USING LINKED LIST.................................................................................................291
TREES........................................................................................................................294
20.1 INTRODUCTION.............................................................................................................................294
20.2 BINARY SEARCH TREES................................................................................................................295

11
Quick & Indepth C With Data Structures

20.3 AVL TREES.....................................................................................................................................302

20.3.1 AVL ROTATIONS....................................................................................................................303
20.4 SPACE AND TIME COMPLEXITIES................................................................................................310
20.5 FULL BINARY TREES.....................................................................................................................311
20.6 PERFECT BINARY TREES...............................................................................................................312
20.7 COMPLETE BINARY TREES............................................................................................................312
SORTING AND SEARCHING.......................................................................................313
21.1 INTRODUCTION.............................................................................................................................313
21.2 QUICKSORT...................................................................................................................................313
21.3 MERGE SORT.................................................................................................................................318
21.4 BINARY HEAP.................................................................................................................................325
21.5 HEAP SORT....................................................................................................................................326
21.6 INSERTION SORT...........................................................................................................................329
21.7 BUBBLE SORT................................................................................................................................330
21.8 SELECTION SORT..........................................................................................................................332
21.9 BINARY SEARCH TREE..................................................................................................................334
21.10 SPACE AND TIME COMPLEXITIES..............................................................................................335
21.11 SEARCHING.................................................................................................................................335
21.11.1 LINEAR SEARCH.................................................................................................................335
21.11.2 BINARY SEARCH.................................................................................................................335
21.11.3 BINARY SEARCH TREE.......................................................................................................337
21.11.4 JUMP SEARCH....................................................................................................................337
C LIBRARY FUNCTIONS.............................................................................................340

12
 INTRODUCTION TO C

INTRODUCTION TO C

[CHAPTER-1]

'C' seems to be an odd name for a programming language. On first thought, it feels it is part of a version
sequence of a programming language. You will not be totally wrong in assuming this. C is an offspring
of the 'Basic Combined Programming Language' (BCPL) called 'B' in short. 'B' was developed at 'Bell
Labs' in 1969 by Ken Thompson and Dennis Ritchie. B was designed for recursive, non-numeric,
machine independent applications, such as system and language software. 'C' was developed by Dennis
Ritchie at 'Bell Labs' between 1969 and 1973 as a successor to the 'B' language. It was originally
developed to re-implement the Unix operating system. It has since become one of the most widely used
programming languages of all time. C has been standardized by the American National Standards
Institute (ANSI) since 1989 (as ANSI C) and subsequently by the International Organization for
Standardization (ISO). C compilers are available across all major architectures and operating systems.
Many languages like C++, D, Java, JavaScript, C#, Objective-C, Perl, PHP, Python have directly or
indirectly borrowed extensively from 'C'. Almost all major operating systems like MS-DOS, Windows,
Mac, Linux, Android, iOS have components written in C or C++ (a successor to C).

1.1 IMPORTANCE OF C
'C' is a robust language which provides a perfect balance between performance and development ease.
It was designed to be compiled using a relatively straightforward compiler, to provide low-level access
to memory, to provide language constructs that map efficiently to machine instructions, and to require
minimal run-time support. Therefore, C was useful for many applications that had formerly been coded
in assembly language, for example in system programming. The C compiler combines the capabilities of
an assembly language with the features of a high level language, making it well suited for writing both
system software and business packages. C being a compiler and so close to assembly language, is
many times faster than most other languages. Unlike assembly language, C is highly portable. A
program written for one system architecture or an operating system can be ported to another after no
or very little modifications and re-compiling for the new system.

C follows the concept of structured programming requiring the user to think of the problem in terms of
function modules or blocks. A proper collection of these modules would make a complete program.
Another important feature of C is its ability for extensions. A C program is basically a collection of
functions that are supported by the C library functions. By continuously adding new functions to the C
library, the language has become more and more powerful and programming task simple.

1.2 WRITING PROGRAMS

C consists of five components - system library files, system header files, user library files, user header
files and user code files.

13
Quick & Indepth C With Data Structures

System library files contain many important library functions which perform specific tasks. A function is
a block of code which performs a specific task. A task can be a combination of the four basic tasks:
Input, Output, Storage and Processing. A function may be called by another function (called the caller)
requiring the later (called the callee) to perform a task on its behalf. The callee may accept parameters
or arguments which may act as inputs for performing the task. The callee may also return a value which
is the result of the processing. Though there can be multiple parameters to a function, there can be only
one return value. Coming back to our previous discussion, system library files (like stdio, stdlib etc.)
contain functions which perform such specific tasks.

System header files (like stdio.h, string.h etc.) contain the prototype or declaration of these system
library functions and data types (discussed later). Prototypes or declarations provide information (like
the function name, what are the parameters that the function accepts, what is its return type) regarding
the function. This information not only helps the programmer to understand the requirements of the
function but also helps the compiler verify the correctness of a function call.

User library files are similar to the system library files except that they are custom written by the users
for extending the system library functions or to perform specific tasks not available in the system
libraries. User libraries are created for porting or re-using the same functions across multiple programs
without re-implementing the same across programs.

User header files contain the function declarations and data types to be shared among different user
code files and user libraries.

The user code files define the actual program functionality utilizing the previous four components and
other processing code.

1.3 OUR FIRST PROGRAM

As our first program, lets print a simple message.

#include <stdio.h>
int main()
{
printf(“This is my first C program.”);
return 0; /* Return 0 to indicate successful operation' */
}

When the above program is executed, it will generate the following output on the console:
This is my first C program.

Now, let us take a first look at the program.

The first statement (#include <stdio.h>) tells the compiler to include the system header file 'stdio.h'.
This header file contains the declaration of the predefined system function 'printf' used within the

14
 INTRODUCTION TO C

program. By predefined we mean that the function body has already been written, compiled and linked
with our program at the time of linking.

Take a look at the 'printf' function call. Over here it is accepting one parameter i.e. The string to print.
The prototype of the 'printf' function is declared as 'int printf ( const char * format, ... )'. The information
contained within the parentheses of the printf function is called the argument or parameter of the
function. The printf function returns the number of characters written to the output, which in this case
is 27 (including the spaces and the dot '.').

The 'main()' is a special function used by the C language which tells the computer where the program
starts. Every program must have exactly one main function. If there are more than one main function,
the compiler will fail to understand the beginning of the program and will raise an error.

The empty pair of parentheses following main tells the compiler that the function does not accept any
arguments in this case. Though the main function can accept two arguments, but you can define a main
function which accepts no arguments. The two arguments which it may accept are int and a character
pointer which we will discuss later. Though there is no standard defined as the return type of the main
function, but, it is a good practice to return an int, which refers to an integer. This return value will tell
the caller program or the operating system regarding the success or failure result of the program
execution. The main function like any other function may return nothing, which is marked with the 'C'
keyword 'void'. We will discuss regarding 'C' keywords a little later.

The opening curly brace '{' appearing after the main function name marks the beginning of the function
main() and the closing curly brace '}' marks the end of the function. Braces group a collection of
statements together and for a function mark the start and end of the function. All the statements
between these two braces specify what the function wants to do and is called the function body. Over
here the function body contains two statements where one is the printf function call and the other
('return') is a 'C' reserved keyword. The value following the return statement is the value returned by
the function to its caller indicating the result of the function processing or operation.

The statement enclosed within '/*' and '*/' following the return statement is not processed by the
compiler and is just a comment for the readability and understanding of the program by the developers.
All lines appearing between '/*' and '*/' is ignored by the compiler. A comment can be inserted in any
place within the program where a blank space can occur. It can appear as separate independent lines,
at the start, middle or end of a line.

All processing statements in 'C' should end with a semi-colon (;) mark.
In place of writing 'printf(“This is my first C program.”);', we could have written the statement as
'printf(“This is my\nfirst C program.”);'. The '\n' refers to a new line character i.e. whenever the
compiler encounters it, all characters appearing after '\n' within the printf statement will be displayed in
the next line of the console. So, in this case the output will be

This is my
first C program.

15
Quick & Indepth C With Data Structures

We could have achieved the same result using two separate consecutive 'printf' statements like
printf(“This is my”);
printf(“\nfirst C program.”);

So, let us recollect our previous understanding regarding the program using the following image.

1.4 STRUCTURE OF C PROGRAMS

Documentation and Comment Section

[Contains information regarding your program like what it does, steps to execute the program etc.]
Preprocessor Statements
[Special instructions for the compiler for compiling the program like header file inclusions etc.]
Declaration and Definition Section
[Function prototypes, structure definitions etc.]
Global Declarations
[Global variable declarations i.e. variables declared outside the functions and applicable to the entire
program]
main() Function Section
{
Function Variable Declarations
Executable statements
}
User-Defined Functions

1.5 EXECUTING A 'C' PROGRAM

Execution of a 'C' program involves a series of steps. They are:
• Creating a program: This involves writing the code of the program.
• Compiling the program: In this step, individual code files are compiled into individual object
files.

16
 INTRODUCTION TO C

• Linking the program: In this step, the individual object files are linked together to generate the
executable.
• Executing the program.

1.6 IDENTIFIERS IN C
An identifier is a string of alphanumeric characters that begins with an alphabetic or an underscore
character that is used to represent various programming elements such as variables, functions, arrays,
structures, unions and so on. Identifier is a user-specified word. So, in our example 'printf' too is an
identifier (called function identifier) and all variable names are also identifiers.

17
Quick & Indepth C With Data Structures

DATA TYPES, VARIABLES AND CONSTANTS

[CHAPTER-2]

2.1 INTRODUCTION
For a programming language, data is the information processed or stored by a computer. This data can
be in the form of numbers or characters. The data is processed and information is generated in the form
of some suitable output. This output can be shown to the user on the screen, printed, stored to a disk or
sent over the network. The instructions or steps for processing the data is called a program. These
instructions are written in some specific rules called syntax rules to make them understandable to the
compiler.

2.2 CHARACTER REPRESENTATION

Characters in C are represented using ASCII (American Standard Code for Information Interchange).
ASCII specifies a correspondence between digital bit patterns and character symbols. A specific digital
bit pattern corresponds to a specific character symbol. As an example the digital bit pattern
'01000001' denotes the character 'A' and the bit pattern '01100001' is for the character 'a'.

ASCII reserves the first 32 codes (numbers 0–31 decimal) for control characters: codes originally
intended not to represent printable information, but rather to control devices (such as printers) that
make use of ASCII, or to provide meta-information about data. For example, character 13 represents a
new line i.e. the cursor on the output device moves to the next line on encountering this character.
Codes from 32 to 126 represent printable characters, including letters, numbers and special characters.
ASCII uses only 7 bits out of a byte to represent one character, leaving 1 bit unused.

Though ASCII could represent almost all English characters, it could not do so for all characters from
other languages. ASCII could represent only 95 characters which is insufficient to handle characters
from other languages as they have a larger character set. So, a new character format called Unicode
was defined which represented one character using more than one byte. The number of bytes used by a
Unicode character depends on the compiler and the encoding mechanism used, which may be 1+ byte
(UTF-8 encoding), 2 bytes (UTF-16 encoding) and 4 bytes (UTF-32 encoding). ASCII characters were
incorporated in the Unicode character set as the first 128 symbols. So, the first 128 symbols are the
same in ASCII and Unicode.

18
 DATA TYPES, VARIABLES AND CONSTANTS

2.3 TOKENS IN C
In a C program, the basic element recognized by the compiler is a token or lexical unit. It is the smallest
unit which is not broken further during parsing. There can be six type of tokens in C.

2.4 KEYWORDS
Keywords are special words with fixed meanings which cannot be changed. They are the words which
are already defined in the compiler. They cannot be used as variable names i.e. no identifier can have
the same spelling and case as a C keyword. Native C compiler had 32 keywords defined but many
modern compilers have extended the list with some of their own. As per ANSI C (a standard published
by American National Standards Institute), the list of C keywords are:

19
Quick & Indepth C With Data Structures

auto break case char

const continue default do
double else enum extern
float for goto if
int long register return
short signed sizeof static
struct switch typedef union
unsigned void volatile while

In later versions of C, more keywords were added to the language and they are:
_Alignas _Alignof _Atomic _Bool
_Complex _Generic _Imaginary _Noreturn
_Static_assert _Thread_local _Pragma

2.5 CONSTANTS
Constants refer to fixed values which do not change during the execution of a program. Constants can
be:

2.5.1 INTEGER CONSTANTS

An integer constant consists of a sequence of digits. An integer constant can be represented in decimal,
octal or hexadecimal.

Decimal are base 10 representation and so consists of digits ranging from 0 to 9. They can be preceded
by an optional – or + sign. Example of decimal integer constants are 562, -8217, +838383, 0 and so on.
Spaces, commas or any other non-digit characters are not allowed within the digits.

Octal are base 8 representation and so consists of digits ranging from 0 to 7. They are preceded with a
leading 0 to help the compiler differentiate them from decimals or hexadecimals. Example of octal
integer constants are 026, 0, 015, 0264.

Hexadecimal are base 16 representation and so consists of digits from 0 to 9 and alphabets from A to F
(or a to f). They are preceded with a leading 0x to help the compiler differentiate them from decimals or
octal. Example of hexadecimal integer constants are 0x26, 0xFF, 0x5AFF.

The largest integer constants are machine dependent . It can be 232-1 for a 32-bit representation and

20
 DATA TYPES, VARIABLES AND CONSTANTS

264-1 for a 64-bit representation. This can be increased by appending qualifiers such as U, L and UL to
the constants like 3324334U (unsigned integer constant), 2234443L (long integer constant),
4234324324UL (unsigned long integer constant).

2.5.2 REAL CONSTANTS

Integer constants cannot be used to represent values which can be in the decimals such as 186.56,
176527.762872, -0.881981. These numbers have a whole number followed by a decimal point and then
the fractional part. It is possible to omit digits before or after the decimal point i.e. a value may not
contain the whole number part and another value may not have the fraction part. Such numbers are
called real constants. A real constant may also be expressed in exponential notation. For example, the
value 678.72 may be expressed as 6.7872e2 in exponential notation where e2 refers to multiplication by
100. 'e' refers to 10 and the number following 'e' refers to its power. The general mathematical
representation is:
n
mantissa x 10
Where mantissa can be any real or an integer number. 'n' is an integer with an optional plus or minus
sign. If 'n' is preceded by a minus sign, it means the resultant multiplication with the mantissa will move
the decimal point to the left rather than the right.

The letter 'e' can be represented in lower or upper case.

Examples of such numbers are 727.321e2, 0.72829e4, 0.072828e5, 8382.382e-1.

Real constants are also called floating point constants and are represented in double precision. They
can be made to represent single precision by suffixing them with a 'f' or 'F'.

2.5.3 SINGLE CHARACTER CONSTANTS

A single character constant (or character constant) contains a single character enclosed within a pair
of single quotation marks such as 'A', 'g', '2', '#',' '. The constant '2' is not the same as the number 2.
The character constant '2' denotes the ASCII value 50 (refer to the ASCII table in para 2.2). For
example, the statement
printf(“%c”, 50);
will print the character '2' on the output screen and the statement
printf(“%d”, 'A');
will print the number 65 on the output screen. 65 is the ASCII value of the character 'A'.

Some special character constants are defined in 'C' which change their meaning and ASCII value
representation. Remember, some non-printable values are also present in the ASCII table. These values
carry special meanings. These constants are sometimes called backslash character constants. They
represent only one character although they are represented using two characters (a backslash followed
by a character). The backslash is called an escape character which changes the meaning of the
character following it. The two characters (‘\’ and the character following it) are processed as a single
unit and are evaluated as a single character. If we need to specify only the backslash character, it itself
needs to be escaped by another backslash.

21
Quick & Indepth C With Data Structures

The various backslash character constants are:

Constant Meaning ASCII Value

'\a' An audible alert 7
'\b' Backspace 8
'\f' Form feed 12
'\n' New line (Line feed) 10
'\r' Carriage return 13
'\t' Horizontal tab 9
'\v' Vertical tab 11
'\0' NULL (marks the end of a string, file etc.) 0
'\'' Single quote 39
'\”' Double quote 34
'\?' Question mark 63
'\\' Backslash 92

2.5.4 STRING CONSTANTS

A string constant is a sequence of characters enclosed in double quotes. The characters can be letters,
numbers, space, backslash characters and special characters. Internally, the compiler marks the ending
of the string constant using a special non-printable character called NULL ('\0') represented by the
ASCII value 0 (zero). The character constant 'G' is different from the string constant “G”. A string
constant does not have an equivalent integer value while a character constant has an equivalent integer
value corresponding to its ASCII value.

2.6 IDENTIFIERS
An identifier is a string of alphanumeric characters that begins with an alphabetic character or an
underscore character and is used to represent various programming elements such as variables,
functions, arrays, structures, unions etc. An identifier can also contain an underscore and is a user-
defined name given to a programming element. Though we can start an identifier with an underscore,
but it is discouraged. An identifier can be of any length but two identifiers must be distinguishable in the
first 31 characters. In other words the C compiler uses the first 31 characters of an identifier as the
token. Case of a character within an identifier is important i.e. it is case sensitive. For example, the
identifier 'NAME' is different than the identifier 'name'. Space is not allowed within an identifier. The
identifier should not be the same as a C keyword.

2.7 VARIABLES
Variables are names used to refer to some location in memory which holds a value with which we are
working. They are placeholder for a value. Before a program can utilize memory to store a variable it
must claim the memory needed to store the value for a variable. This is done by declaring the variable.
Before a variable is used the compiler needs to know the number of variables it needs, what they are
going to be named, and how much memory they will need. When managing and working with variables,
it is important to know the type of variables and the size of these types. The type determines the kind of

22
 DATA TYPES, VARIABLES AND CONSTANTS

data we are going to store in the variable, for example integer values (0, 9828, 2827382, ...), character
values ('G', 'M', 'a', '#', ...), real values (728.28, 822329.283302, 62.3738e3, ...) etc. Unlike constants
which cannot be changed during the lifetime of the program, values in variables can be changed
according to the needs of the program. Variable names are governed by identifier naming convention as
stated earlier. Some valid variable names are 'm_dblAverage', 's_fltUnitPrice', 'ulTotal'. Some invalid
variable names are '353dhd', 'edfsd$22&35', '@', '4545'.

2.7.1 VARIABLE DECLARATION

After determining variable names, we must declare the variables before using them anywhere in the
program. This helps the compiler understand the memory requirements of the variable and the kind of
operations to be allowed on them. Declaration helps the compiler understand the following things:
• The variable names.
• The type of data the variable will contain.
• The memory required to hold the value corresponding to the variable.
• The operations to be allowed on the variable.

Variables can also be declared using the optional 'const' keyword which makes the value contained
within the variable constant i.e. the value contained within the variable cannot be changed.

2.8 DATA TYPES

Data types are declarations for memory locations or variables that determine the characteristics of the
data that may be stored and the operations that are permitted using them. ANSI C supports four
classes of data types:
1. Primary (basic or fundamental) data types.
2. Derived data types.
3. Alias data types.
4. void data type.

2.8.1 PRIMARY DATA TYPES

• int: They hold integer values. They are whole numbers and their capacity depends on the
compiler but, must use at-least 16 bits (2 bytes) of internal storage. Generally most modern
compilers use 32 bits (4 bytes) of internal storage to represent an integer value allowing an int
variable to hold values in the range of -2147483648 to +2147483647 (i.e. -231 to +231-1). A
signed integer uses one bit to hold the sign of the value and the rest for holding the magnitude.
There are two kinds of modifiers which can be used to change the properties of the integer data
– Storage modifiers and sign modifiers. Storage modifiers 'short' and 'long' can be used to
change the range of the values stored in a variable whereas the sign modifiers 'signed' and
'unsigned' can be used to change the sign of the value. 'short' modifier halves the storage
requirement of the value and also the range of the value. It makes the data use 16 bits (2 bytes)
of internal storage giving it a range of -32768 to +32767 (i.e. -215 to +215-1). 'long' modifier
doubles the storage requirement and the range of the value but, is machine architecture
dependent. On 32-bit machines the 'long' modifier makes the data use 32 bits (4 bytes) of

23
Quick & Indepth C With Data Structures

internal storage giving it a range of -2147483648 to +2147483647 (i.e. -231 to +231-1). On 64-
bit machines it makes the data use 64 bits (8 bytes) of internal storage giving it a range of
-9223372036854775808 to +9223372036854775807 (i.e. -2 63 to +263-1). As the long modifier
makes the range differ depending on the machine architecture, a new modifier 'long long' was
introduced in some newer compilers which makes the data use exactly 64 bits (8 bytes) of
internal storage regardless of the machine architecture. 'signed' modifier is optional and all
integer values are signed by default unless we explicitly specify 'unsigned' as the modifier.
'unsigned' modifier specifies that the value will only store 0 and positive values.

• char: They hold character type data like 'T', 'Q', 'h', '4', '?' etc. Characters are usually stored in
8 bits (1 byte) of internal storage. This is sufficient to hold all character representations in the
English language, but is insufficient to hold all characters of many languages other than
English. So, a new character type called 'Unicode' (wide char denoted using 'wchar_t') was
defined which stores the value in 1 to 4 bytes (depending on the encoding used) of internal
storage (refer section 2.2). We will discuss more about Unicode in our later chapters. Just like
'int', the sign modifiers 'signed' and 'unsigned' may also be applied to 'char'. While signed 'char'
can hold values ranging from -128 to 127, unsigned 'char' can hold values in the range of 0 to
255.

• float: They hold real numbers like 3444.3839, 89393.33, 83.3443e2 etc. They are stored in 32
bits (4 bytes) of internal storage and have 6 digits of precision. No size or sign modifier can be
applied to the float data type and all float values are always signed.

• double: When the accuracy provided by float is not sufficient, the type 'double' can be used.
They are stored in 64 bits (8 bytes) of internal storage and have 14 digits of precision. They
hold the same type of data (real) as float but with double the precision, hence the name
'double'. Their precision can be further extended to use 80 bits (10 bytes) of internal storage by
using the size modifier 'long'. The other modifiers 'short', 'signed' and 'unsigned' cannot be
applied to the double data type. They are always signed.

• Data types specific to C99 and later: In the year 1999, some modifications were made to the C
programming language to make it more robust and extend it with more features. It helped
utilize the current hardware better and achieve better program optimization. It introduced new
data types which were not available in earlier versions. The data types that were introduced
were float_t, double_t, float _Complex, double _Complex, long double _Complex and _Bool.
◦ float_t and double_t: They correspond to the types used for the intermediate results of
floating-point expressions when FLT_EVAL_METHOD is 0, 1, or 2. FLT_EVAL_METHOD
specifies the precision in which all floating point arithmetic operations other than
assignment and cast are done. When FLT_EVAL_METHOD is:
▪ 0 (zero): float_t and double_t are equivalent to float and double, respectively.
▪ 1: Both float_t and double_t are equivalent to double.
▪ 2: Both float_t and double_t are equivalent to long double.

24
 DATA TYPES, VARIABLES AND CONSTANTS

◦ float _Complex, double _Complex and long double _Complex: When the floating point
types are followed by the '_Complex' keyword, they help store and process mathematical
complex numbers. Complex numbers are numbers which are of the form 'a + bi' where 'a'
and 'b' are real numbers and 'i' is an imaginary unit.
◦ _Bool: _Bool is a type which is capable of holding the two values 1 and 0, where 1 signifies
'true' and 0 signifies 'false'. _Bool has been type defined as 'bool' in the header file
'stdbool.h'. So, for convenience we can also write 'bool' in place of '_Bool', if we include the
header file 'stdbool.h'.

2.8.2 SIZE AND RANGE OF PRIMARY DATA TYPES

Type Size (bytes) Range
char/signed char 1 -128 to 127.
unsigned char 1 0 to 255.
wchar_t 2 0 to 65535 [Unicode characters].
short/short int/signed 2 -32768 to +32767.
short/signed short int
unsigned short/ 2 0 to 65535.
unsigned short int
int/signed/signed int 2 or 4 -32768 to +32767 when represented using 2 bytes..
-2147483648 to +2147483647 when represented using 4 bytes.
unsigned/unsigned int 2 or 4 0 to 65535 when represented using 2 bytes.
0 to 4294967295 when represented using 4 bytes.
long/long int/signed 4 or 8 -2147483648 to +2147483647 when represented using 4 bytes.
long/signed long int -9223372036854775808 to +9223372036854775807 when
represented using 8 bytes.
unsigned long/ 4 or 8 0 to 4294967295 when represented using 4 bytes.
unsigned long int 0 to 18446744073709551615 when represented using 8 bytes.
long long/long long 8 -9223372036854775808 to +9223372036854775807.
int/signed long long/
signed long long int /
_int64
unsigned long long/ 8 0 to 18446744073709551615.
unsigned long long int/
unsigned _int64
float 4 -3.4e+38 to +3.4e+38.
double 8 -1.7e+308 to +1.7e+308.
long double 10 Typically -3.4e+4932 to +1.1e+4932.

Note: The data types with multiple size options in the ‘Size (bytes)’ column (ex: 2 or 4), are compiler
dependent and so, the number of bytes they occupy depend on the compiler with which we are working.

25
Quick & Indepth C With Data Structures

2.8.3 DERIVED DATA TYPES

These data types are either formed using a combination of primary data types or they are used to
represent the primary data types in a different way. They extend the scope of C language. The most
common are pointers, arrays, union and structures (struct). We will read about them in later chapters.

2.8.4 ALIAS DATA TYPES

C allows type-definition which lets the programmer define an identifier that will represent an existing
data type. In other words, we may give a new name (alias) to an existing data type. This alias can later
be used to declare variables. There are two types of alias data types:
• Data types created using the typedef keyword.
• Data types created using the enum keyword.

'typedef' is a reserved keyword in the C programming languages. It is used to create an alias name for
another data type. As such, it is often used to simplify the syntax of declaring data structures consisting
of struct and union types, but can also be used to provide specific descriptive type names for the
primary data types, arrays and pointers. The syntax of a typedef statement is:

typedef type identifier;

where 'type' refers to an existing data type and 'identifier' refers to the new typedef (alias) name given
to the data type. The new name remains absolutely identical to the data type which it is type-defining.
'typedef' cannot create a new type but provides just an alias name for an existing type. Some examples
of typedef statements and their uses are:

typedef int categories;

categories PrimaryID;
categories SubIDs[100];

Another alias data types is defined using the keyword 'enum' (enumerated in short). An enumeration
consists of a set of named integer constants. An enumeration type declaration gives a name (optional)
and defines a set of named integer identifiers representing values which the enumerated name can
hold. It is defined as:
Example 1: enum <name>
{
identifier1,
identifier2,
identifier3,
…
};

In the above example, the compiler internally provides an integer value of 0 to 'identifier1', 1 to

26
 DATA TYPES, VARIABLES AND CONSTANTS

'identifier2', 2 to 'identifier3' and so on. So, if we do not provide any specific values, the enumerated
identifiers will obtain values starting from 0 and increment sequentially.
'enum' can also be defined in a way so that the enumerated identifiers can attain certain specific integer
values like:
Example 2: enum <name>
{
identifier1 = 10,
identifier2 = 11,
identifier3 = 12,
identifier4 = 20,
identifier5 = 21,
…
};
In the above example, the compiler internally provides an integer value of 10 to ' identifier1', 11 to
'identifier2', 12 to 'identifier3', 20 to ‘identifier4’ and so on.

We can then declare variables using the enumerated name in the following manner:
enum <name> v1;
enum <name> v2;
where the variables 'v1' and 'v2' can only hold values identified using the identifiers (identifier1,
identifier2, ...) as mentioned in the enumeration.

Usage example:
enum DayOfTheWeek
{
Monday,
Tuesday,
Wednesday,
Thursday,
Friday,
Saturday,
Sunday
};

enum DayOfTheWeek StartDay;

enum DayOfTheWeek EndDay;

StartDay = Wednesday; /* StartDay is assigned Wednesday (integer value 2). */

EndDay = StartDay; /* EndDay will also have the value Wednesday. */
if (EndDay == Wednesday)
{
EndDay = Sunday; /* Now, EndDay gets assigned Sunday (integer value 6). */
}

27
Quick & Indepth C With Data Structures

2.8.5 void DATA TYPE

The data type void actually refers to no value or no specific data type. A function may return void which
means that it does not return anything. A function may also accept void as its parameter signifying that
it does not accept any parameters or arguments. Similarly, we can have void pointers (discussed in later
chapters), which mean that they point to memory locations whose data type is not known.
Example: void PrintVal(int I);
int GetVal(void);
void ShowWelcomeMessage(void);

2.9 STORAGE CLASSES

A storage class defines the scope (visibility) and life-time of variables within a C Program. They precede
the type that they modify. They tell the compiler where to store a variable, how to store the variable,
what is the initial value of the variable and life time of the variable. Consider the following example:
int g_Val1;
void function1(void)
{
int iVal2;
}

int main(void)
{
int iVal3;
int iVal2;
function1();
return 0;
}

The variable 'g_Val1' which has been defined outside all functions is called the global variable. It is
accessible across all the functions within the source file. It need not be declared in any function to be
accessible.

The variable 'iVal2' is a local variable to the function 'function1' and is available only to this function. All
other functions do not have access to this variable. Similarly, the variable 'iVal3' is local to the the
function 'main' and is accessible only within the function 'main'. We can see that the variable ' iVal2' has
also been declared within the 'main' function. But, the 'iVal2' of the function 'function1' is different than
the 'iVal2' of the function 'main'. Both can have different values, are stored in separate internal memory
locations and are accessible only to the respective functions.

C programs can have four storage class specifiers. They are:

28
 DATA TYPES, VARIABLES AND CONSTANTS

Storage Class Meaning Default Value

auto Local variable known only to the function in which it is declared. It Garbage
is the default. (Uninitialized)
Usage Example: int iCount; auto double dblPrice;
static The static storage class instructs the compiler to keep a local 0
variable in existence during the life-time of the program instead of
creating and destroying it each time it comes in and goes out of
scope. Making local variables static enables them to maintain their
values between function calls. The static modifier may also be
applied to global variables upon which it causes that variable's
scope to be restricted not just to a function but to the entire file in
which it is declared.
Usage Example: static char iVal;
extern The extern keyword is used before a variable to inform the compiler 0
that this variable is declared in some other source file of the same
program. Global variables declared in one source file can be
accessed in another using the keyword extern. For example, an
integer 'iVal' declared in the source file 'file1.c' can be accessed in
another file 'file2.c' using the extern keyword.
Usage Example: extern long lTotal;
register Local variables which are stored in the CPU registers. For some Garbage
high performance temporary operations, we may declare register (Uninitialized)
variables. As they are stored in registers, a variable's maximum
size is limited to the register size (usually one word). Only few
variables can be placed inside register.
Usage Example: register short Counter;

2.10 DEFINING SYMBOLIC CONSTANTS AND MACROS

We often require to use some constant values within the program. These values may need to be used
repeatedly across the source files of the program. As an example, we may require to use 1.414 as the
square root value of 2. We may be using it repeatedly within the program. So, to save ourselves from
errors we may define it using a name, and use the defined name across all places which require that
value. The general form of defining is:
#define symbolic-name value

where symbolic-name is the name given to the value. As an example, we may use it as:
#define SQUARE_ROOT 1.414

The defined names are called symbolic names or symbolic constants. When the program is compiled,
these symbolic names are replaced by their respective defined values by the compiler before
performing the compilation. As an example, the symbolic name 'SQUARE_ROOT' is replaced by 1.414,
wherever that name is encountered within the program. These statements are called preprocessor
statements as they are processed before the compilation begins. The preprocessor statements begin

29
Quick & Indepth C With Data Structures

with a '#' symbol, and do not end with a semicolon. The defined value may represent a numeric
constant, a character constant, a string constant or even an expression. Though there is no restriction
regarding the case of the symbolic name, it is considered a convention to use all UPPERCASE
characters. This helps in distinguishing them from normal variable names which are written using lower
case or mixed (both lower an upper) case characters.
The '#define' statement can also define an expression to be used in place of the symbolic-name
wherever it is used. The expression can be something like:
#define INCREMENT(x) ++x

Wherever the compiler finds the symbolic name 'INCREMENT', it replaces the name with the expression
corresponding to the name specified in the '#define' statement. This kind of statements are also called
macros. A Macro is typically an abbreviated name given to a piece of code or a value. A macro is a code
which has been given a name and whenever the name is used, it is replaced by the contents of the
macro. The argument(s) to the macro ('x' in our example) is replaced by the actual argument(s)
specified during the macro call. There can be more than one argument to the macro. So, we may have
something like:
/* Header Files */
#include <stdio.h>
/* Defined Values */
#define INCREMENT(x) ++x

void main(void)
{
/* Variable declaration and initialization. */
int i = 10;

/* Now, print the incremented value. */

printf("\nAfter increment i is: %d.\n\n", INCREMENT(i));
}

In the above program, the macro call 'INCREMENT' within the 'printf' statement will be replaced by the
expression corresponding to the macro and 'i' (specified within the call) will replace 'x' within the macro
expression. So, the macro call 'INCREMENT(i)' will be replaced by the expression '++i'. The program will
print the following output:

After increment i is: 11.

The following rules apply to the #define statement.

• '#' must be the first character of the line, though some modern compilers allow spaces before
the '#' symbol.
• No space should be present between the '#' and the 'define' word.
• At-least one blank space is required between '#define' and the symbolic name, and between the
symbolic name and the constant value.
• Preprocessor statements (like #define) must not end with a semi-colon.
• The data type of the constant is determined by the compiler depending on its value. The
'#define' statement does not specify any data type.

30
 DATA TYPES, VARIABLES AND CONSTANTS

• After a symbolic name has been defined, it cannot be re-defined anywhere within the source file
(unless undefined). Also, NO value can be assigned to a symbolic name anywhere else within the
program (like SQUARE_ROOT = 1.4).
• '#define' statements can appear anywhere within the program but before they are referenced.
By convention they are placed at the beginning of a source or header file.
• Separate preprocessor statements (like #define) must appear in separate lines. Neither multiple
values can be defined for a symbolic name, nor multiple definitions can appear in a single line.
So, the following statements are invalid.
#define SQUARE_ROOT 1.414 1.414213
#define SQUARE_ROOT 1.414 PI 3.141
#define SQUARE_ROOT 1.414, PI 3.141

Now, let us discuss the advantages of using the '#define' statements. There are primarily two
advantages of using '#define' statements and they are:
• Easy modifications: If we use the square root value of 2 as 1.414 directly across multiple
locations within the program, it will require an immense amount of work in case we later decide
to change the value. Say, we later decide to increase the precision of the value from the existing
1.414 to 1.414213, we will need to make changes across the entire program wherever that
value gets used. If any location is missed, the program may generate erroneous output.

• Easy understanding: If we had used the constant value 1.414 directly within a program
location, another programmer may not understand the origin of that value. Multiple
programmers may work on the same program, or a program may be developed by one
programmer but, maintained by another. When a new programmer starts working on the
program, (s)he may not understand the origin or the reason for using that specific constant
value. If a descriptive name is defined for the value and that name is used across the program,
(s)he will immediately understand the reason and the origin of the value.

2.11 DECLARING A VARIABLE AS A CONSTANT

We may not want a value to be defined using the '#define' statement as we cannot specify its data type.
We may require a value to be of a certain specific data type. So, for values like these we may use the
'const' keyword. The 'const' keyword is used as a qualifier during variable declaration. The variable
must be declared and initialized in the same statement when using the 'const' qualifier. This constant
variable can be used as the right hand value of any assignment statement and also provided as
argument to any function. No value can be assigned to the variable. If a constant variable is declared
within a function, its value cannot be modified anywhere within the function, and every time the function
gets invoked, the constant variable gets the same value with which it is initialized. If a constant variable
is declared within a source file, its value cannot be changed anywhere within the entire source file
(including all its functions). Some examples of the 'const' qualifier are:
const float fltPI = 3.141;
const int iPrice = 130;

31
Quick & Indepth C With Data Structures

2.12 VOLATILE VARIABLES

We can have another qualifier called 'volatile' which gets applied during variable declaration. It tells the
compiler that the value of the variable may change at any time, and even without any action being taken
by the code. If not properly used, the implications of this behaviour may be quite serious. In real life we
should rarely need to use the 'volatile' qualifier. An example use of the 'volatile' qualifier is given below.
volatile int iValue;

If a variable is declared as volatile, it can, not only be modified by external factors (like processes,
hardware operations etc.), but also be changed by its own program as well. To prevent the value from
getting changed by the own program and yet allowing external forces to change the value, we may add
the 'const' qualifier along with the 'volatile' qualifier as below.
volatile const int iValue = 56;

2.13 DATA OVERFLOW AND UNDERFLOW

When programming we need to be very careful regarding assigning values to variable which are larger
or smaller than the variable's capacity. Data overflow occurs when we try to assign a value to a variable
which is larger or smaller than the variable's capacity (as per its data type). When overflow occurs, the
value in the variable gets rolled over (wrap around). C does not provide any warnings for data overflow.
It only generates wrong results which may make the program to fail. Underflow can occur when the true
result of a floating point operation is smaller in magnitude (that is, closer to zero) than the smallest
value that can be represented in the target data type.

2.14 SAMPLE PROGRAM

Calculation of area of circles.

Lets try to write a program where we will calculate the area of multiple circles. The radius of the circle
will start from 5 and will increment sequentially till 15.
/*******************************************************/
/* Area of circles */
/* Copyright Information */
/*******************************************************/
/* Header Files */
#include <stdio.h>

/* Defined Values */
#define PI 3.141592
#define MIN_RADIUS 5
#define MAX_RADIUS 15

void main(void)
{
/* Variable Declaration & Initialization */
int iRadius = MIN_RADIUS;
double dblArea = 0;

while (iRadius <= MAX_RADIUS)

{

32
 DATA TYPES, VARIABLES AND CONSTANTS

dblArea = PI * iRadius * iRadius;

printf("\nThe area of the circle with radius %d is %f.", iRadius, dblArea);
iRadius = iRadius + 1;
}
}

Output
The area of the circle with radius 5 is 78.539800.
The area of the circle with radius 6 is 113.097312.
The area of the circle with radius 7 is 153.938008.
The area of the circle with radius 8 is 201.061888.
The area of the circle with radius 9 is 254.468952.
The area of the circle with radius 10 is 314.159200.
The area of the circle with radius 11 is 380.132632.
The area of the circle with radius 12 is 452.389248.
The area of the circle with radius 13 is 530.929048.
The area of the circle with radius 14 is 615.752032.
The area of the circle with radius 15 is 706.858200.

33
Quick & Indepth C With Data Structures

OPERATORS AND OPERATIONS

[CHAPTER-3]

3.1 INTRODUCTION
An operator is a symbol that operates on a value or variable. It is used to perform certain mathematical
or logical operations on data and variables. C language supports a rich set of built-in operators. The
operators in C can be classified into the following categories.
1. Arithmetic operators
2. Increment and decrement operators
3. Assignment operator
4. Bitwise operators
5. Relational operators
6. Logical operators
7. Conditional (Ternary) operator
8. Special operators
The operations performed using the operators are also called expressions.

3.2 ARITHMETIC OPERATORS

C provides operators for all the basic arithmetic operations. The arithmetic operators are:

Operator Meaning
+ Addition of two numbers or unary plus.
- Subtraction of two numbers or unary minus.
* Multiplication of two numbers.
/ Division of two numbers.
% Remainder after division (Modulo division).

Arithmetic operations are always performed between two numbers. The result of an arithmetic
operation between two numbers is always in the larger data type of the two numbers involved in the
operation. By larger data type we mean the data type which is capable of holding a larger value or
higher precision. The following rules are applied in the order given:
• If one of the number is a long double, the other number too is promoted to long double and the
result will be in long double.
• If one of the number is a double, the other number too is promoted to double and the result will
be in double.
• If one of the number is a float, the other number too is promoted to float and the result will be
in float.
• If one of the number is long int, the other number too is promoted to long int and the result will
be in long int.

34
 OPERATORS AND OPERATIONS

• If one of the number is int, the other number too is promoted to int and the result will be in int.
• If one of the number is short, the other number too is promoted to short and the result will be in
short.
• The result will be in char, if both the numbers are in char.
The arithmetic operation between two integer is called integer arithmetic and the expression is called
integer expression. Integer arithmetic always yields an integer value. The integer arithmetic between
two integer values 'x' and 'y' where 'x' has a value of 19 and 'y' has a value of 8, will yield the following
results.

x – y = 11
x + y = 27
x * y = 152
x / y = 2 (decimal part truncated as the result will be in integer)
x % y = 3 (remainder of the division)

During division between two integer (int) numbers, the decimal part is truncated as the result of two
integer (int) numbers will always be an integer (int). If we need the remainder, we may use the modulo
(%) operator.

During modulo division, the sign of the result is the sign of the first operand in the operation. For
example:

-19 % 8 = -3
-19 % -8 = -3
19 % -8 = 3
19 % 8 = 3

An arithmetic operation between two real numbers is called real arithmetic operation. A real operand
may be in decimal or exponential notation. The result of such an operation is a real number. The modulo
operator (%) cannot be used with real operands.

When in an arithmetic operation, one of the operand is a real number and the other an integer, the
operation is called mixed mode arithmetic operation. The result of such an operation is a real number.
For example:

39 / 10.0 = 3.9

3.3 INCREMENT AND DECREMENT OPERATORS

Increment and decrement operators are unary operators that add or subtract one from their operand,
respectively. They are:
++ and --

The increment operator is written as ++ and the decrement operator is written as --. The increment

35
Quick & Indepth C With Data Structures

operator increases the value of its operand by 1 and the decrement operator decreases the value by 1.
The operands must have an arithmetic or pointer data type. They take the following forms:
++count or count++ (Same as: count = count + 1)
--count or count-- (Same as: count = count -1)

Though, ++count and count++ do the same thing, they behave a bit differently when they are put in the
right hand side of an assignment statement.
Consider the following statements:
int count = 10;
int index = count++;

In the above statements, the value of 'count' will be 11 whereas the value of 'index' will be 10. So, in
other words the value of 'count' is assigned to 'index' before performing the increment. The operation
'count++' is called post-increment, as the increment is done after the assignment.
Now consider the following statements:
int count = 10;
int index = ++count;

In the above statements, the value of 'count' will be 11 and the value of 'index' will also be 11. So, in
other words the value of 'count' is incremented before assigning it to 'index'. The operation '++count' is
called pre-increment, as the increment is done before the value is assigned.

In simple words, in case of post-increment, the operator first assigns the value to the variable on the
left and then increments the operand. In case of pre-increment, the operator first increments the
operand and then the result is assigned to the variable on the left. Similar behavior occurs for post-
decrement and pre-decrement operations.

The increment and decrement operators may be used in complex statements. Say, we have two
variables 'x' with a value of 15 and 'y' with a value of 10. The execution breakup corresponding to some
expressions are as follows:

Expression Execution Breakup Result

m = x++ * ++y; 1) ++y (y is incremented) m: 165, x: 16, y: 11
2) m = x * y (x is multiplied with y and the value assigned to
m)
3) x++ (x is incremented)
m = x++ + ++y; 1) ++y (y is incremented) m: 26, x: 16, y: 11
2) m = x + y (x is added with y and the value assigned to m)
3) x++ (x is incremented)
m = x++ + ++y * 2; 1) ++y (y is incremented) m: 37, x: 16, y: 11
2) m = x + y * 2 (y is multiplied by 2, added with x and the
result assigned to m)
3) x++ (x is incremented)

36
 OPERATORS AND OPERATIONS

Expression Execution Breakup Result

m = x-- + ++y * 2; 1) ++y (y is incremented) m: 37, x: 14, y: 11
2) m = x + y * 2 (y is multiplied by 2, added with x and the
result assigned to m)
3) x-- (x is decremented)
m = ++x + y-- * 2; 1) ++x (x is incremented) m: 36, x: 16, y: 9
2) m = x + y * 2 (y is multiplied by 2, added with x and the
result assigned to m)
3) y-- (y is decremented)

3.4 ASSIGNMENT OPERATORS

Assignment operators are used to assign the result of an expression to a variable. Apart from the usual
assignment statement 'variable = expression;', C provides another kind of assignment statement in the
following form:
var op= expression;
where 'var' is for variable and 'op' is for an arithmetic operator. The operator ' op=' is called the
shorthand assignment operator.

Consider the following examples:

index += 5;
index *= count – 2;

The first statement is similar to the following statement:

index = index + 5;

And the second statement is similar to the following statement:

index = index * (count – 2);

Shorthand assignment operator examples

Expression with shorthand operator Similar expression with simple assignment operator
count -= 1; count = count - 1;
count /= (j + 1); count = count / (j + 1);
count %= j; count = count % j;
count += k * 5; count = count + (k * 5);
count *= k + 3; count = count * (k + 3);

3.5 BITWISE OPERATORS

C is one of the few languages which allow manipulations at the bit level. We can do manipulations or
comparisons at the bit level using the bitwise operators. Bitwise operators can be used only on 'char'
and 'int' data types. There are five bitwise operators and they are:

37
Quick & Indepth C With Data Structures

Operator Meaning
& bitwise AND
| bitwise INCLUSIVE OR (OR)
^ bitwise EXCLUSIVE OR (XOR)
~ bitwise NOT
<< SHIFT LEFT
>> SHIFT RIGHT

Now, lets describe all the above bitwise operators with some examples. Say we have two unsigned char
variables, 'x' with a value of 77 and 'y' with a value of 55. In binary format 'x' will be represented as
'01001101' and 'y' will be represented as '00110111'. Now, lets explain what happens when we
perform the above bitwise operations on these two variables.
• & (bitwise AND): The bitwise 'AND' operator compares each bit of the first operand to the
corresponding bit of the second operand. If both the bits are 1, the corresponding result bit is
set to 1. Otherwise, the corresponding result bit is set to 0. In the following operation the value
of 'm' after the 'AND' operation will be 5.
unsigned char m = x & y;

• | (bitwise INCLUSIVE OR): The bitwise 'INCLUSIVE OR' or simply 'OR' operator compares each
bit of the first operand to the corresponding bit of the second operand. If either bit is 1, the
corresponding result bit is set to 1. Otherwise, the corresponding result bit is set to 0. In the
following operation the value of 'm' after the 'OR' operation will be 127.
unsigned char m = x | y;

• ^ (bitwise EXCLUSIVE OR): The bitwise 'EXCLUSIVE OR' or simply 'XOR' operator compares
each bit of the first operand to the corresponding bit of the second operand. If one bit is 0 and
the other bit is 1, the corresponding result bit is set to 1. Otherwise, the corresponding result
bit is set to 0. In the following operation the value of 'm' after the 'EXCLUSIVE OR' operation will
be 122.

38
 OPERATORS AND OPERATIONS

unsigned char m = x ^ y;

• ~ (bitwise NOT): The bitwise 'NOT' operator looks at the binary representation of the operand
and does a bitwise negation operation on it. Any bit that is a 1 in the operand becomes 0 in the
result and any bit that is a 0 becomes 1 in the result. When the ~ operator acts on an operand it
returns a value of the same data type as the operand.
unsigned char m = ~x;

• << (SHIFT LEFT): The bitwise 'LEFT SHIFT' operator shifts each bit in its left-hand operand to
the left by the number of positions indicated by the right-hand operand. The left most bits
(numbering equal to the value of the second/right-hand operand) are simply dropped-off in the
result. For the number of bits shifted towards the left, equal number of 0s (zeroes) are inserted
from the right side. In the following example, 3 left-most bits are dropped-off and 3 zeroes are
inserted from the right-hand side in the final output in 'm'.
unsigned char m = x << 3;

• >> (SHIFT RIGHT): The bitwise 'RIGHT SHIFT' operator shifts each bit in its left-hand operand
to the right by the number of positions indicated by the right-hand operand. The right most bits
(numbering equal to the value of the second/right-hand operand) are simply dropped-off in the
result. For the number of bits shifted towards the right, equal number of 0s (zeroes) are
inserted from the left side. In the following example, 2 right-most bits are dropped-off and 2
zeroes are inserted from the left-hand side in the final output in 'm'.
unsigned char m = x >> 2;

39
Quick & Indepth C With Data Structures

Bitwise operators can be part of some complex expressions like the ones given in the following table.
Expression Simplified Result
m = x | y >> 3; x | (y >> 3); m: 79, x: 77, y: 55
m = x | y & x ^ y; x | ((y & x) ^ y); m: 127, x: 77, y: 55
m = x++ | --y & x << 3; 1) --y (y is decremented) m: 109, x: 78, y: 54
2) m = x | (y & (x << 3));
3) x++ (x is incremented)

3.6 RELATIONAL OPERATORS

The relational operators compare values to one another. When we decide to do certain operations
depending on some conditions, these conditions are generally represented using relational operators.
For example, we may compare the price of two products and decide to purchase the product with the
lower price. This comparison may be done using the '<' or '>' relational operators. The expression
represented using a relational operator is termed as a relational expression. One relational expression is
generally composed of one relational operator and two operands (values) on both sides of the operator.
The result of a relational expression is either 'true' or 'false' where false is denoted by '0' (zero) and
true is non-zero. In C, true has the value of 1, but may be compiler specific. For example:
int x = 5;
int y = 10;

x < y → Expression result is true.

x > y → Expression result is false.

Different relational operators available in C are:

Operator Meaning
< Operand at the left of the operator is less than the right-hand operand
> Left-hand operand is greater than the right-hand operand
<= Left-hand operand is less than or equal to the right-hand operand
>= Left-hand operand is greater than or equal to the right-hand operand
== Left-hand operand is equal to the right-hand operand
!= Left-hand operand is not equal to the right-hand operand

Multiple relational expressions may be combined using logical operators (explained later) to create one
single expression.

Remember: In places where an expression requires a relational operator (ex: within an if condition,
conditional operator or loop condition) but none has been specified, then by default the expression is
compared for inequality with 0 (zero). So, the statement ‘if (x)’ is the same as specifying ‘if (x != 0)’. We
will learn more about ‘if conditions’ and ‘loop conditions’ in our later chapters.

Assume we have two variables x and y with values 5 and 10 respectively. On this assumption, the result
of some relational expressions are given in the next table.

40
 OPERATORS AND OPERATIONS

Expression Result
x<y true
x <= y true
x != y true
x == y false
x>3 true
x >= 7 false

3.7 LOGICAL OPERATORS

Logical operators are mainly used to control program flow and make decisions. There are three logical
operators in C.
&& logical AND
|| logical OR
! logical NOT

Logical operators '&&' and '||' are used for combining multiple relational expressions into a single
expression. The result of the individual relational expressions are combined to form a single overall
result. Multiple conditions are evaluated to make a decision. An expression which combines two or more
relational expressions is termed logical expression or compound relational expression. Like the relational
expressions, logical expressions also yield the result as 'true' or 'false'.
• && (AND): The '&&' operator combines two expressions as below.
expression1 && expression2
The logical AND operator (&&) returns true (1) if result of both the expressions (expression1
and expression2) are true (1) and returns false (0) otherwise.
Logical AND is evaluated from left to right. The second expression is evaluated only if the result
of the first expression evaluates to true (non zero). This evaluation eliminates needless
evaluation of the second expression when the result of the first expression is false. For two
variables x and y with values 5 and 10 respectively, we can have the following expressions.
Examples:
x < y && x != 2 → Evaluates to true
x <= 5 && y >= 5 → Evaluates to true
x > y && y == 10 → Evaluates to false as expression1 is false.

• || (OR): The '||' operator combines two expressions as below.

expression1 || expression2
The logical OR operator (||) returns true (1) if result of any of the expressions (expression1 or
expression2) is true (1), and returns false (0) if result of both the expressions are false (0).
Logical OR is evaluated from left to right. The second expression is evaluated only if the result of
the first expression evaluates to false (zero). This is simply because, if result of the first
expression is true, the overall result of the entire logical expression also becomes true. This
eliminates needless evaluation of the second expression when the result of the first expression
is already true. For two variables x and y with values 5 and 10 respectively, we can have the
following expressions.

41
Quick & Indepth C With Data Structures

Examples:
x < y || x != 5 → Evaluates to true as expression1 is true.
x < 5 || y >= 5 → Evaluates to true as expression2 is true.
x > y || y != 10 → Evaluates to false as both expressions are false.

• ! (NOT): The '!' operator performs logical negation on an expression. It is stated as:
!expression
If result of the expression is true (1), NOT operator converts the overall result to false (0). If
result of the expression is false (0), the overall result becomes (1). For a variable x with value of
5, we can have the following expressions.
Examples:
!(x == 5) → Evaluates to false
!(x > 5) → Evaluates to true

3.8 CONDITIONAL (TERNARY) OPERATOR

The conditional operator pair (? :) takes three operands and so is also called ternary operator. It has
the following syntax:
expression1 ? expression2 : expression3

• The expression1 is evaluated first.

• If the result of expression1 is true (non zero), expression2 is evaluated.
• If the result of expression1 is false (zero), expression3 is evaluated.

The result of the conditional operator is the result of whichever expression is evaluated — the second or
the third. We should make sure that the data type of the result of expression2 and expression3 are the
same. If not so, implicit type conversions occur which may not result in the desired output. Consider
three variables x, y and z where 'x' and 'y' have values 5 and 10 respectively. We will assign the result of
some conditional expressions to 'z'. Some examples are given below.

Expression 'z' Explanation

z = (x > y) ? x : y; 10 As expression1 '(x > y)' is false, the result of
expression3 'y' is assigned to 'z'.
z = (x < y) ? (x * 3) : (y * 5); 15 As expression1 is true, the result of expression2 '(x *
3)' is assigned to 'z'.
z = (x != y) ? 1 : 2; 1 As expression1 is true, the result of expression2 '1' is
assigned to 'z'.
z = (x < y) ? (x > 0) ? x : 1 : (y > 0) ? y : 1; 5 Here, both expression2 and expression3 have inner
expressions. As expression1 '(x < y)' is true,
expression2 '(x > 0) ? x : 1' is evaluated and its result
is assigned to 'z'. The result of expression2 is the
value of 'x' as '(x > 0)' is true. The entire expression
is similar to:
z = (x < y) ? ((x > 0) ? x : 1) : ((y > 0) ? y : 1);

42
 OPERATORS AND OPERATIONS

Expression 'z' Explanation

The entire expression actually tries to assign the
smaller number out of 'x' and 'y' to 'z'. But, the value
of 'z' cannot be lower than 1.

3.9 SPECIAL OPERATORS

There are four special operators in 'C' and they are:
• Comma operator (,)
• sizeof operator
• Pointer operators (& and *)
• Member selection operators (. and →)

Here we will discuss only the comma (,) and sizeof operators. The pointer operators (& and *) and
member selection operators (. and →) will be discussed in subsequent chapters.

3.9.1 COMMA OPERATOR

The comma operator links related expressions together. A comma separated list of expressions are
evaluated from left to right and the value of the right-most expression is the final value of the combined
expression.
Let us understand this with an example.
z = (x = 5, y = 10, x * y);

The expression within the first bracket is evaluated from left to right. First, 5 is assigned to 'x' and then
10 is assigned to 'y'. At last, 'x' is multiplied by 'y' and the result is assigned to 'z'. So, 50 (5 * 10) is
ultimately assigned to 'z'. Since, comma operator has the lowest operator precedence i.e. it is the last
one to be evaluated, the parentheses (first brackets) are necessary. If not done so, the assignment
operator will take precedence and 'z' will acquire a value of 5 instead of 50.

3.9.2 sizeof OPERATOR

The sizeof is a compile time operator which means that the expression is evaluated during the
compilation time. sizeof is also a 'C' keyword. When used with an operand, it returns the number of
bytes the operand occupies in memory. The operand may be a variable, a constant or a data type. It can
also be used to determine the total bytes used by arrays, structures and unions. Some examples of
sizeof operator usage are:
z = sizeof(y);
z = sizeof(double);
z = sizeof(36L);

3.10 ARITHMETIC EXPRESSIONS

Arithmetic expressions are combination of variables, constants, operators and parentheses. The values
contained within the variables are used for the arithmetic calculations. The parentheses are used for

43
Quick & Indepth C With Data Structures

prioritizing or sequencing the arithmetic operations. We can make operations with lower precedence
executed before others by enclosing them in parentheses. The result of an arithmetic expression can be
assigned to a variable using the assignment operator. When done so, the previous value of the variable
is replaced by the newly assigned value. Assume we have two variables 'x' and 'y' with values 5 and 10.
Some possible arithmetic expressions are:
z = x * 5 + y;
z = x * (5 + y);
z = x * 5 + (y*3);

The result of the second statement will be different from the first statement. The first bracket or
parentheses in the second statement make the addition operation get performed before the product.
So, when the first statement assigns 35 '(5 * 5) + 10' to 'z', the second statement assigns 75 '5 * (5 +
10)'. Notice that in the third statement we didn't specify any space within '(y*3)'. The space between
operands and operators is optional and is recommended only for better readability.

Arithmetic expressions without parentheses are evaluated from left to right but, depends on the
precedence of the operators. The precedence of the arithmetic operators are divided into two
categories – high precedence and low precedence. The arithmetic operators with high precedence are
* (multiplication), / (Division) and % (Modulus). All these three operators have the same precedence
and are evaluated from left to right i.e. the operator on the left of an expression is evaluated first
followed by the one on the right. The arithmetic operators with low precedence are + (Addition), -
(Subtraction). These two operators have the same precedence and are evaluated from left to right. As
stated earlier, the parentheses can be used to change the precedence or order of the evaluation. Lets
look at one more example.
z = 15 + x * y / 2 % 10 – 3;

The order of operations in the above example are:

Step count Step Result
1 x*y → 50 [5 * 10]
2 50 / 2 → 25 [Result of #1 divided by 2]
3 25 % 10 →5 [Result of #2 modulus 10]
4 15 + 5 → 20 [15 plus result of #3]
5 20 – 3 → 17 [Result of #4 minus 3]
6 17 is assigned to z.

While working with arithmetic expressions, we need to be careful regarding few things.
• While assigning real numbers to integer data type variables, the mantissa will be lost. So for
example, trying to assign '6.75' to a variable 'z' of type 'int', only '6' will be assigned to 'z', as 'z'
being an 'int' type is incapable of holding the mantissa.
• Result of an expression depends on the data type of the variables involved in the expression.
This can be explained by a simple example. Assume, we have a variable 'x' of type int with value
15. The division of this variable by an integer constant or variable will yield a result as int, even
if we assign the result to a float or double.

44
 OPERATORS AND OPERATIONS

double z = x / 2; → 'z' will be assigned a value of 7.

To evaluate the expression in the data type of the target variable, we need to typecast the right-
hand side variables to the data type of the target variable 'z'. We need to remember the
precision rules as stated in section 3.2. All the below statements will assign 7.500000 to 'z'.
double z = (double) x / 2;
double z = x / (double) 2;
double z = (double) x / (double) 2;
• We should make sure that we do not divide by 0 (zero). Division by 0, will make the program
terminate abnormally. Care should be taken to test the denominator for 0 before doing any
division.
• We should also be careful of overflow and underflow errors during arithmetic operations. We
should make sure that the operands are of correct type and are capable of holding the result of
the operation. Overflow and underflow errors generate wrong results. For more details on
overflow and underflow errors please refer to section 2.13.
• We should be careful regarding the precedence and associativity of the operators involved in
the expression. If ignored, we may get completely unexpected results.

3.11 TYPE CONVERSIONS

On many occasions we need to use variables or constants of mixed data types within arithmetic
expressions. We may also need to assign a variable or constant of a different data type to another
variable. We may also require to pass variable or constant of a different data type as argument to a
function instead of the one the function accepts. On all these occasions, the data type of the source
variable or constant needs to be converted to the data type of the target variable. Typecasting is a way
to make a variable or constant of one type, such as an int, act like another type, such as a char, for one
single operation. There are two ways in which we can do type conversions.
• Implicit (Automatic) type conversion: If one of the operand in an expression is of a 'lower' type
(lower capacity) than the other, the lower type operand is automatically converted by the
compiler to the 'higher' type (higher capacity) operand. This conversion is done automatically
by the compiler and we need not specify anything explicitly. For the implicit conversion rules,
please refer to section 3.2. The final result of the expression is converted to the type of the
variable to which the result is getting assigned i.e. the variable on the left-hand side of the
assignment operator. However, we should remember that conversion from real number (float or
double) to integer (char, short int, int, long int) results in truncation of the mantissa. Also,
conversion from a higher type to a lower type results in dropping of excess higher order bits
resulting in reduced overall precision or capacity. Some examples of implicit type conversion
are:

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
char ch = 'A';
int idx = 35;

45
Quick & Indepth C With Data Structures

double val = 0;

/* char is automatically converted to type int before the addition. */

idx = idx + ch;

/* int is automatically converted to double

i.e. the data type of 10.5 before the addition. */
val = idx + 10.5;

/* 110.500000 will be printed. */

printf("\nval: %lf.\n", val);

return 0;
}

• Explicit type conversion: There are instances when we need to force a type conversion in a way
which is different from what the automatic conversion would have achieved. This process is also
called type casting and it is user defined. Here the user can type cast the result to convert it to
a particular data type. This is better explained using an example. Say, in a class we need to find
the ratio of the number of girls to the boys. Though, the number of girls and boys are whole
integer values, we will need the ratio to be in decimals. When calculating the ratio, we will need
to typecast the girl-number and boy-number to double to get the ratio in double.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
unsigned int uiNumGirls = 26;
unsigned int uiNumBoys = 31;
double dblRatio = 0;

/* Both uiNumGirls and uiNumBoys are typecasted to

double before the division. This results in the
division operation to retain the fractional part. */
dblRatio = (double)uiNumGirls / (double)uiNumBoys;

/* 0.838710 will be printed. */

printf("\nRatio: %lf.\n", dblRatio);

return 0;
}

The general form of typecast is:

(data-type) expression;

The expression can be an expression, variable or constant. An expression should be typecasted

before using its value in further arthmetic operations for getting correct results. Some examples
of typecast are given in the next table.

Example Behavior
int x = (int) 56.6; 56 is assigned to 'x' after truncation of the fractional part.
double x = (int) 37.5 / 6; 37.5 is converted to int resulting in its value as 37, which in turn is divided

46
 OPERATORS AND OPERATIONS

Example Behavior
by the integer 6. The division is between two integers (int) and so yields
the result in integer (int). 37 is divided by 6 to get the result as 6 after
truncating the fractional part. The final result of 6 is assigned to the
variable of type double.
float x = (float) 37 / 6; 37 is converted to float resulting in its value as 37.000000, which in turn is
divided by the integer 6. But, 6 is automatically (implicitly) converted to
type float (i.e. the higher type) before the division is performed. The
division is between two floating point values. The final result 6.166667 is
assigned to the variable 'x' of type float.

3.12 OPERATOR PRECEDENCE AND ASSOCIATIVITY

Precedence determines the order of evaluation of the operators. Every operator has a precedence
associated with it. Precedence determines how an expression involving more than one operator is
evaluated. If an expression contains multiple operators, the operator with the higher precedence is
evaluated first. The operators of the same level of precedence is either evaluated from left to right or
right to left depending on the precedence level. This is called the associativity of an operator.
Associativity is used when two operators of same precedence appear in an expression. When evaluating
an arithmetic expression, we should be very careful regarding the precedence and associativity of the
operators involved in the expression. Many a times, programmers either ignore and wrongly interpret
the precedence and associativity resulting in incorrect results. Consider the following statement.
int x = 10 + 25 * 3 – 15 / 5;

In the above statement '25 * 3' is evaluated first followed by '15 / 5'. On doing so, our expression gets
simplified to '10 + 75 – 3'. Now, as the '+' (addition) and '-' (subtraction) are at the same level of
precedence, their associativity property comes into play resulting in the expression to be evaluated from
left to right. So, '10 + 75' is evaluated first resulting in the final result of 82.

If we have confusion regarding the precedence and associativity of the operators, an easy way is to
enclose the operations which need to be evaluated first within parentheses. We can also have nested
parentheses i.e. one pair of parentheses stated within another. When done so, the innermost
parentheses are evaluated first followed by the outer ones. Examples:
(x + (y – z)) / 2;
(x % 5) – y + 10;
(x << 3) + y * z;

A complete list of the operators and their associativity are given in the next table.

Precedence Operator Description Associativity

1 () Function call, and within arithmetic operations Left-to-right
[] Array element reference
. Structure or union member access
-> Structure or union member access using pointer

47
Quick & Indepth C With Data Structures

Precedence Operator Description Associativity

2 + Unary plus Right-to-left
– Unary minus
++ Increment
–– Decrement
! Logical NOT
~ Bitwise NOT
* Pointer indirection (dereference)
& Address-of
sizeof Size-of
(data type) Type cast
3 * Multiplication Left-to-right
/ Division
% Modulus
4 + Addition Left-to-right
– Subtraction
5 << Bitwise LEFT SHIFT Left-to-right
>> Bitwise RIGHT SHIFT
6 < Less than Left-to-right
<= Less than or equal to
> Greater than
>= Greater than or equal to
7 == Relational equality Left-to-right
!= Relational inequality
8 & Bitwise AND Left-to-right
9 ^ Bitwise XOR Left-to-right
10 | Bitwise OR Left-to-right
11 && Logical AND Left-to-right
12 || Logical OR Left-to-right
13 ?: Conditional expression (Ternary operator) Right-to-left
14 = Simple assignment Right-to-left
*= Assignment by product
/= Assignment by division
%= Assignment by remainder (modulus)
+= Assignment by addition
-= Assignment by subtraction
&= Assignment by bitwise AND
^= Assignment by bitwise XOR
|= Assignment by bitwise OR
<<= Assignment by bitwise LEFT SHIFT
>>= Assignment by bitwise RIGHT SHIFT
15 , Comma operator Left-to-right

48
 OPERATORS AND OPERATIONS

3.13 SAMPLE PROGRAMS

1. Swap the values of two variables using a temporary variable.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
int iVar1 = 5;
int iVar2 = 10;
int iTemp = 0;

/* Swap */
iTemp = iVar1;
iVar1 = iVar2;
iVar2 = iTemp;

/* Print the swapped values. */

printf("After swap, iVar1: %d, iVar2: %d.\n", iVar1, iVar2);

return 0;
}

Output:
After swap, iVar1: 10, iVar2: 5.

2. Swap the values of two variables without using a temporary variable.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
int iVar1 = 5;
int iVar2 = 10;

/* Swap */
iVar1 += iVar2;
iVar2 = iVar1 – iVar2;
iVar1 = iVar1 – iVar2;

/* Print the swapped values. */

printf("iVar1: %d, iVar2: %d.\n", iVar1, iVar2);

return 0;
}

Output:
After swap, iVar1: 10, iVar2: 5.

3. Circular shift bits 3 places to the left. All bits are moved 3 places to the left and the highest 3
bits move to the lowest 3 places in the final result.
How do we do it?

49
Quick & Indepth C With Data Structures

We first left-shift the data 3 places to the left, and then we 'OR' it with the highest 3-bits of the
data. This internally gets done in 3 steps.
a) When we left-shift the data 3 places, the highest 3 bits drop-off and 0 ( zeroes) get
inserted in the lowest 3 bit positions. If our value is 187 which has a binary
representation of '10111011', our left-shifted value becomes '11011000'.
b) To get the highest 3 bits to the lowest 3 positions we right-shift the data equal to 5
places (8 - 3) i.e. the number of bits in a byte minus the number of high order bits we
want to put in the lower positions. When we right-shift the value '10111011' 5 places,
it becomes '00000101'.
c) The result obtained in step (a) gets 'OR'ed to the result obtained in step (b) resulting in
the final value of '11011101' (221).

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
/* Binary representation of ch is 10111011.*/
unsigned char ch = 187;
/* cShift refers to the number of places to shift.
It should not be greater than 8. */
unsigned char cShift = 3;

/* Circular shift the bits 'cShift' places to the left.

This can also be done in 3 separate steps as given below.
unsigned char temp1 = ch << 3;
unsigned char temp2 = ch >> (8 – 3);
ch = temp1 | temp2;
The above 3 statements will also achieve the same result. */
ch = (ch << 3) | (ch >> (8 – 3));

/* Print the shifted value. */

printf("Circular shift of ch yields: %d.\n", ch);

return 0;
}

Output:
Circular shift of ch yields: 221.

4. The price of a package of 25 chocolates is $14. If the chocolates are sold separately, an extra 6
cents are charged for every chocolate. Find the price of 65 such chocolates assuming that 50
chocolates will be sold in 2 packages and the rest 15 will be sold separately.
/* Header Files */
#include <stdio.h>

/* Defined Values */
#define PACKAGEQTY 25
#define PACKAGEPRICE 14.0f
#define EXTRACHARGE 0.06f

int main(void)
{
/* Variable Declaration & Initialization */

50
 OPERATORS AND OPERATIONS

int iNumChocToPurchase = 65;

int iCompletePackage = 0;
int iNumChocSoldSeparate = 0;
double dblRatePerChoc = 0;
double dblFinalPrice = 0;

/* Calculate the price per chocolate when sold separately. */

dblRatePerChoc = PACKAGEPRICE / PACKAGEQTY + EXTRACHARGE;

/* Calculate the number of complete packages we will get at the

price of 'PACKAGEPRICE' for 'PACKAGEQTY' chocolates. */
iCompletePackage = iNumChocToPurchase / (int) PACKAGEQTY;

/* Calculate the number of chocolates which we need to purchase separately*/

iNumChocSoldSeparate = iNumChocToPurchase - iCompletePackage * PACKAGEQTY;

/* Now, calculate the final price for the total purchase. */

dblFinalPrice = iCompletePackage * PACKAGEPRICE + iNumChocSoldSeparate *
dblRatePerChoc;

/* Print the final price. */

printf("Amount needed to purchase %d chocolates is $%.2lf.\n",
iNumChocToPurchase, dblFinalPrice);

return 0;
}

Output:
Amount needed to purchase 65 chocolates is $37.30.

51
Quick & Indepth C With Data Structures

INPUT, OUTPUT AND FORMAT SPECIFIERS

[CHAPTER-4]

4.1 INTRODUCTION
As discussed in section 1.3, we can use the 'printf' function to print results on the standard output
device. Similarly, we have a function 'scanf' for accepting inputs from the user. In C, all input and output
operations are performed through function calls such as 'printf' and 'scanf'. These functions are
collectively called standard input and output functions and are part of the C library. Every program
which uses these standard input and output functions must include the header file 'stdio.h' at the
beginning of source file or header, using the following statement.
#include <stdio.h>
where 'stdio' stands for standard input and output.

The statement '#include <stdio.h>' tells the compiler to search for a file named 'stdio.h' and place that
entire file's contents at that location within the program. The content of the header file becomes part of
the source code when it is compiled. Some modern intelligent compilers do not include the header file
contents, if nothing (functions, defines etc.) from the header file is referenced within the program.

Another header file which we are going to discuss in this chapter is 'ctype.h'. This header declares a set
of functions to classify and transform individual characters.

4.2 READING INPUT

4.2.1 SIMPLE INPUT FUNCTIONS

There are multiple functions in C which can be used to read keyboard input. Each function is targeted
towards a specific way of accepting the input. The simplest functions are the ones which accept a single
character. There are three such functions.
• int getch(void): This function is specific to MS-DOS supporting compilers and is not available
on most Unix and Linux targeted compilers. This function has been deprecated and is
superseded by the function 'int _getch(void)' which has the same behavior as 'getch'. It waits
for a keyboard input and returns the ASCII value of the character entered. The function does
not echo the entered character to the console screen. Consider the following example usage.
char chVar = getch();

• int getche(void): Like 'getch', this function too is specific to MS-DOS supporting compilers and
is not available on most Unix and Linux targeted compilers. This function has been deprecated
due to some security concerns and is superseded by the function 'int _getche(void)' which has
the same behavior as 'getche'. It waits for a keyboard input and returns the ASCII value of the
character entered. The function echoes the entered character to the console screen. Consider

52
 INPUT, OUTPUT AND FORMAT SPECIFIERS

an example usage.
char chVar;
chVar = getche();

• int getchar(void): This function gets a character (an unsigned char) from standard input (stdin)
device, typically the keyboard. To indicate an error, it returns EOF (-1). This function is available
in all C compilers. The function has the following form of declaration and usage:
Declaration → int getchar( void );
Usage → Variable_Name = getchar();

Where 'Variable_Name' is a variable generally of type 'char'. When this statement is

encountered, the compiler waits until a key is pressed. It returns the entered character which
gets assigned to the variable 'Variable_Name'. We may also have 'Variable_Name' of 'int' data-
type, if we are looking to work with the ASCII value of the entered character.

'getchar' function can also be used to read a sequence of inputs from the user within a loop
statement. We will know more about the loop statements in later chapters.

The 'getchar' function accepts any key available in the keyboard buffer including the RETURN
(ENTER) key and the TAB key. Consider the following example usage.
char chVar = ' ';
printf(“Please enter a character”);
chVar = getchar();

Unlike the above functions which accept only a single input character, we can have an entire string
(including spaces within the string) accepted as input using the 'gets' function. The function takes the
following form.
char * gets(char * buffer);
Where 'buffer' is a character array which accepts the entered string. An array is a sequence of values of
the same data type stored in contiguous memory locations. We will read about arrays in later chapters.
The function accepts the pointer to the buffer where the string needs to be stored and returns the
pointer to the starting address of the 'buffer'. We will read about pointers too in later chapters. We
should avoid using this function as it does not check for buffer overflows and may write past the buffer.

4.2.2 FORMATTED INPUT

Till now we learned about accepting only a specific type of keyboard input – either single character or a
character string. In C, we can also accept formatted data from the keyboard. Formatted input refers to
an input data formatted to a particular data type. Assume, we accept four types of input: 627,
52611.26, 'G' and “Football”. These four data are of four different types. The value 627 is an integer
and can be held in the data type 'int'. The second value 52611.26 is a real number and must be held in
the data types 'float' or 'double'. The third value 'G' is a single character and can be held in the data
type 'char'. The fourth value “Football” is a sequence of characters called a string. A NULL (ASCII value
0 [zero]) character marks the end of the string.

53
Quick & Indepth C With Data Structures

We can accept formatted data from the standard input device using the ' scanf' function. The function is
declared in the header file 'stdio.h' and has the following form.
int scanf(“format string”, arg1, arg2, …. argn);

The 'format string' specifies the data formats to which the input needs to be accepted. The argument
list specify the memory address of the variables in which the accepted input needs to be stored. The
number of arguments must equal the number of format specifiers in the 'format string'. The format
string may also contain spaces, tabs and new lines which are ignored. Spaces are used for better
readability. In our next section, let us see discuss format specifiers.

4.3 FORMAT SPECIFIERS

Each format specifier is of the form:
% [flags] [field_width] [.precision] [length_modifier] conversion_character
where components in brackets [] are optional.

Format specifiers tell the compiler regarding the type of value to be read or printed. This can be
explained using a simple example. If we want to print the value '1', the format specifier enables the
compiler to understand whether we want to print the character '1' or the number 1. The character '1'
has an ASCII value of 49 whereas if represented as a number it will simply be 1.
Format specifiers must start with a percentage (%) operator.

'conversion_character' specifies the type of the value i.e. int, char, float etc. They can be:
Conversion character Meaning
c Single character.
d, i Signed decimal integer.
u Unsigned decimal integer.
o Octal integer.
x, X Hexadecimal integer.
f Signed floating point value.
e, E Signed floating point value in scientific notation.
g, G Signed floating point or signed scientific notation, whichever is shorter.
s A character string. A string is a sequence of characters ending in NULL.
p Pointer address. Displays the argument as an address in hexadecimal digits.
n Nothing is printed.
The corresponding argument must be a pointer to a signed int.
The number of characters written so far is stored in this pointed location.
% A % followed by another % character will write a single % to the stream.

The 'flags' are applicable for number values only. They can be:
Flag Meaning
- (minus) The output is left justified (aligned) in its field, not right justified which is the default.
+ (plus) Signed numbers will always be printed with a leading sign (+ or -). The default is to
print the '–' (minus) sign but not the '+' plus.

54
 INPUT, OUTPUT AND FORMAT SPECIFIERS

Flag Meaning
space Prepends a space for positive signed numeric types. This flag is ignored if the '+' flag
exists. The default does not prepend anything in front of positive numbers.
0 (zero) When the 'field_width' option is specified, prepends zeros for numeric types. The
default prepends spaces.
# Alternate form:
For g and G types, trailing zeros are not removed.
For f, e, E, g, G types, the output always contains a decimal point.
For o, x, X types, the text 0, 0x, 0X, respectively, is prepended to non-zero numbers.

The 'field_width' specifies a minimum number of characters to output. If the converted argument has
fewer characters than the field width, it will be padded on the left (or right, if left adjustment has been
requested) to make up the field width. The padding character is normally ' ' (space), but can be '0' if the
zero padding flag (0) is present. It does not cause truncation of oversized fields. It is typically used to
pad fixed width fields as tabulated output. Consider the following example.
int iValue = 5;
char szName[ ] = “John”;
printf("Value is %02d", iValue);
printf("Value is %2d", iValue);
printf(“Name is %6s”, szName);

The above three statements will print the following output in a single line (we are showing in three
separate lines for understandability and readability purposes):
Value is 05
Value is 5
Name is John

The '.precision' field specifies the maximum limit on the output, depending on the particular formatting
type. For floating point numeric types, it specifies the number of digits to the right of the decimal point
that the output should be rounded to. For the string type, it limits the number of characters that should
be printed, after which the string is truncated. Consider the following example.
float fltValue = 2.35435;
printf("Value is %.2f", fltValue);
The above statement will print:
Value is 2.35

'length_modifier' tells the compiler regarding the size or capacity of the data. For example the format
specifier '%hu' signifies short unsigned integer. The different length modifiers are:
Length modifier Meaning
h The capacity of the number is limited to short. It can be used with d, i, u, o, x and
X conversion characters.
l The capacity of the number is extended to long. It can be used with d, i, u, o, x , X
and f conversion characters. When specified with 'f', it signifies the data type
'double'.

55
Quick & Indepth C With Data Structures

Length modifier Meaning

ll The capacity of the number is extended to long long. It can be used with d, i, u, o,
x and X conversion characters.
L Long double. It can be used with e, E, f, g and G conversion characters.

Some example usage of length modifiers are '%hu', '%hd', '%lu', '%ld', '%lf', '%llu', '%lld', '%Lf' etc.

4.3.1 FORMATTED INPUT [EXAMPLES]

As we discussed, we can accept formatted input using the 'scanf' function. For our understanding, let us
take six variables of different data types. We will then accept values from the user into these variables in
separate 'scanf' functions.

/* Header Files */
#include <stdio.h>

int main(void)
{
char chValue = ' ';
short shValue = 0;
int iValue = 0;
unsigned int uValue = 0;
double dblValue = 0;
char szName[101] = "";

printf("Please enter a single character [char]: ");

scanf("%c", &chValue);

printf("\nPlease enter a number between -32768 to +32767 [short signed int]: ");
scanf("%hd", &shValue);

printf("\nPlease enter a number between -2147483648 to +2147483647 "

"[signed int]: ");
scanf("%d", &iValue);

printf("\nPlease enter a number between 0 to +4294967295 [unsigned int]: ");

scanf("%u", &uValue);

printf("\nPlease enter a real number [float]: ");

scanf("%lf", &dblValue);

printf("\nPlease enter a name within 100 characters [string]: ");

scanf("%s", szName);

printf("\n\nThe values entered are: %c, %hd, %d, %u, %lf and %s.", chValue,
shValue, iValue, uValue, dblValue, szName);

return 0;
}

4.4 GENERATING OUTPUT

4.4.1 SIMPLE OUTPUT FUNCTIONS

In previous sections, we have seen the various input functions available in C. Similarly, we have output

56
 INPUT, OUTPUT AND FORMAT SPECIFIERS

functions which are classified into two types – Functions for non formatted output and functions for
formatted output. Let us first discuss about the simpler functions for producing non formatted output.
• Similar to the input 'getchar' function, we have an output function 'putchar' which produces a
single character on the standard output device (stdout). This function is available in all C
compilers. The function has the following form of declaration and usage:
Declaration → int putchar( int character );
Usage → putchar( Variable_Name );

Where 'Variable_Name' is a variable generally of type 'char'. Few example usage are given
below.
char chVar = 'W';
putchar( chVar );
putchar( 'O' );
putchar( chVar );
putchar( '!' );
putchar( '\n' );

The above statements will print the following word on the screen:
WOW!

• Like the 'gets' input function we have the 'puts' output function which produces a character
string on the standard output device (stdout). It also appends a newline character ('\n') at the
end of the string. The function has the following form:
int puts ( const char * str );

The function begins copying from the address specified (str) until it reaches the terminating
NULL character ('\0'). This terminating NULL character is not written to the output. We will read
about variable addresses (pointers) in later chapters.

4.4.2 FORMATTED OUTPUT

We have been using the 'printf' function in our various examples for printing captions and results. We
should generate the output in such a way which is understandable to the user. We use the 'printf'
function not only to give instructions but also present the output. The 'printf' function is declared in the
header file 'stdio.h' and has the following form:
int printf(“format string”, arg1, arg2, …. argn);

The function writes the C string pointed by the 'format string' to the standard output (stdout). If format
includes format specifiers, the additional arguments following the format string are formatted and
inserted in the resulting string replacing their respective specifiers. We can also include backslash
character constants within the format string (refer section 2.4.3). The number of format specifiers in
the format string should equal the count of the arguments following the format string. Also, the format
specifiers should match the data type of the corresponding arguments. The format specifiers should
follow the rules as stated in section 4.3. Consider the following code as an example.

57
Quick & Indepth C With Data Structures

/* Header Files */
#include <stdio.h>

int main(void)
{
printf("Characters:\t\t\t %c, %c \n", 'a', 65);
printf("Decimals:\t\t\t %d, %ld \n", 2007, 83839394L);
printf("Integer preceding with blanks:\t %10d \n", 2007);
printf("Integer preceding with zeroes:\t %010d \n", 2007);
printf("Width adjusted:\t\t\t %*d \n", 5, 100);
printf("Different radices:\t\t %d, %x, %o, %#x, %#o \n", 2007, 2007, 2007, 2007,
2007);
printf("Floating point numbers:\t\t %4.2f, %+.0e, %E \n", 3.1416, 3.1416, 3.1416);
printf("A string:\t\t\t %s \n", "C Language");

return 0;
}

The above code will yield the following output:

Characters: a, A
Decimals: 2007, 83839394
Integer preceding with blanks: 2007
Integer preceding with zeroes: 0000002007
Width adjusted: 100
Different radices: 2007, 7d7, 3727, 0x7d7, 03727
Floating point numbers: 3.14, +3e+00, 3.141600E+00
A string: C Language

4.5 SAMPLE PROGRAMS

1. Accept two numbers from the user and calculate the average (mean) of the two.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
unsigned int uNumber1 = 0;
unsigned int uNumber2 = 0;
double dblAverage = 0;

/* Accept the first number. */

printf("Enter the first number: ");
scanf("%u", &uNumber1);

/* Accept the second number. */

printf("Enter the second number: ");
scanf("%u", &uNumber2);

/* Calculate the average. We are typecasting the numbers to a higher data type
(double) because addition of the two numbers could create an overflow
condition, if the entered numbers are of very high value. Also, typecasting
helps us get the result in fraction which we intend to. */
dblAverage = ((double)uNumber1 + (double)uNumber2) / 2;

/* Print the result. */

58
 INPUT, OUTPUT AND FORMAT SPECIFIERS

printf("\nThe average of the two numbers %u and %u is %.02lf.\n", uNumber1,

uNumber2, dblAverage);

return 0;
}

2. Accept the length of two sides (legs) of a right-angled triangle and calculate the hypotenuse
using the Pythagorean theorem.
/* Header Files */
#include <stdio.h>
#include <math.h>

int main(void)
{
/* Variable Declaration & Initialization */
double dblSide1 = 0;
double dblSide2 = 0;
double dblHypotenuse = 0;

/* Accept the length of the first side. */

printf("Enter the length of the first side: ");
scanf("%lf", &dblSide1);

/* Accept the length of the second side. */

printf("Enter the length of the second side: ");
scanf("%lf", &dblSide2);

/* Calculate the hypotenuse.

It can also be calculated as
dblHypotenuse = sqrt(dblSide1 * dblSide1 + dblSide2 * dblSide2); */
dblHypotenuse = hypot(dblSide1, dblSide2);

/* Print the result. */

printf("\nThe hypotenuse of the triangle is %.02lf.\n", dblHypotenuse);

return 0;
}

In the above program, we have used the math function 'hypot' to calculate the hypotenuse of
the triangle. Instead of it, we could have also used the following statement to calculate the
hypotenuse.
dblHypotenuse = sqrt(dblSide1 * dblSide1 + dblSide2 * dblSide2);
The 'hypot' function takes the length of two sides of the triangle as 'double' and returns the
hypotenuse of the triangle also as double. The 'sqrt' function takes a number as double and
returns its square root value. Both these functions are declared in the header file 'math.h' and
so we have included that header at the start of our program.

3. Accept the horizontal and vertical sides of a rectangle and calculate its area and perimeter.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
double dblSideHorizontal = 0;

59
Quick & Indepth C With Data Structures

double dblSideVertical = 0;
double dblArea = 0;
double dblPerimeter = 0;

/* Accept the length of the first side. */

printf("Enter the length of the horizontal side [1 - 9999]: ");
scanf("%lf", &dblSideHorizontal);

/* Accept the length of the second side. */

printf("Enter the length of the vertical side [1 - 9999]: ");
scanf("%lf", &dblSideVertical);

/* Calculate the area. */

dblArea = dblSideHorizontal * dblSideVertical;

/* Calculate the perimeter. */

dblPerimeter = (dblSideHorizontal + dblSideVertical) * 2;

/* Print the result. */

printf("\nThe area of the rectangle is %.02lf and the perimeter is %.02lf.\n",
dblArea, dblPerimeter);

return 0;
}

60
 ARRAYS

ARRAYS

[CHAPTER-5]

5.1 INTRODUCTION
An array is a collection of data items, all of the same type, accessed using a common name. An array is
used to store a collection of data, but, it is often more useful to think an array as a collection of
variables of the same type. For example, we may define an integer (int) array with the name 'MathMark'
which stores the marks obtained in Mathematics by all the students of a specific class or standard. In
the definition, the array name is followed by the array capacity within square brackets. Assuming we
have 20 students in the class, the array will be defined as:
int MathMark[20];

The general form of an array declaration is:

Data-Type ArrayName[Dimension1-Size][Dimension2-Size]...;

An individual value within the array is accessed using the array name followed by the index number or
subscript. In C, array indexes or subscripts start from 0 (zero) and continue to 1 less than the array
capacity. So in our example, the indexes will start from 0 and continue till 19 (20 – 1). The marks of the
first student will be at the index location 0 and that of the last will be at index location 19. To store or
access the marks obtained by the 15th student in the class, we need to use the following statements:
Store → MathMark[14] = <Value>;
Access → MathMark[14]

While the complete set of values is referred to as an array, the individual values within the array are
called elements. Arrays can be of any data type. Arrays are stored in memory unless we persist them to
disk using any of the file management functions which we will read in later chapters. As they are stored
in memory, they will be lost when the program terminates unless we save the array contents to disk.

Arrays are created using contiguous memory locations. The lowest address corresponds to the first
element of the array and the highest address belongs to the last element. A one-dimensional array will
be represented as:

Assuming 'int' takes 4 bytes and our array starts at an imaginary memory location of 100, the first
element (MathMark[0]) will be located at the memory location 100, followed by 104 for 'MathMark[1]',
108 for 'MathMark[2]' and so on.

61
Quick & Indepth C With Data Structures

Arrays help us to write concise and efficient programs enabling us to store and access values within
loops. An array can be of single or multi-dimension. A dimension is a direction in which you can vary the
specification of an array's elements. A one-dimensional array is like a list whereas a two dimensional
array is like a table. The C language places no limits on the number of dimensions in an array, though
specific implementations may vary.

5.2 ONE-DIMENSIONAL ARRAYS

A one-dimensional array is like a list which has only one subscript and is also known as single-
subscripted array. The example which we discussed in section 5.1 is one-dimensional array. The
general form of a one-dimensional array is:
Data-Type ArrayName[Array-Size];

The values are assigned to a one-dimensional array in the following manner:

MathMark[0] = 88;
MathMark[1] = 81;
MathMark[2] = 92;
.
.
.
MathMark[19] = 68;

Array elements can be used just like any other C variable. Any operation which can be performed on a
variable of the same data type as that of the array, can also be performed on an array element. For
example, if 'Numbers' is an array of type 'int' with capacity of 10 elements:
int iValue = Numbers[5] – 3;
Numbers[2] = Numbers[4] + Numbers[7];
Numbers[0] = Numbers[3] * 3;
Numbers[3] = 6 * 8;
Numbers[8] += 17;

C does not perform any bound checking on arrays. If care is not taken, we may overshoot the array
index and thus try to access a memory location outside the array-size. This will result in a run time error
and lead to program crash. So, we need to be careful and do bound checking before accessing any
array element.

5.2.1 INITIALIZATION OF ONE-DIMENSIONAL ARRAYS

Just like oridinary variables, array elements can be initialized during declaration time. The general form
of initialization of a one-dimensional array is:
Data-Type ArrayName[Array-Size] = { comma separated list of values };

Where the comma separated values are assigned to the array elements sequentially in the provided

62
 ARRAYS

order. Say, if we specify 3 initialization values, then the first value gets assigned to the first array
element; the second value gets assigned to the second array element; and the third value to the third
array element. If the array-size is more than the number of values specified, the remaining elements will
automatically be initialized to 0 (zero).
int rgiNumbers[5] = { 62, 81, 23 };
In the above example, the value 62 is assigned to 'rgiNumbers[0]', 81 is assigned to 'rgiNumbers[1]'
and 23 is assigned to 'rgiNumbers[2]'. The other two elements 'rgiNumbers[3]' and 'rgiNumbers[4]' are
automatically assigned 0 (zero) by the compiler.

When we are initializing the array during declaration, the array-size becomes optional. If omitted, the
size of the array equals the number of initialization values specified. So, if we have provided three
comma separated values during initialization, the array will have three elements. Consider the following
example.
float rgfltTotal[] = { 13.5, 16.7, 89, 107.6, 70.0 };
The above statement creates a float array of 5 elements, as we have provided 5 initialization values.

Initialization of arrays in C has two drawbacks.

1. We cannot initialize specific elements. Values are assigned sequentially from the first element.
2. Large number of array elements cannot be initialized in one go. We need to use separate loop
statements to initialize the elements.

Character arrays can be initialized in the same way as below.

char rgchInitial[] = { 'S', 'T', 'A', 'R' };
Please note that the above statement declares an array called 'rgchInitial' with 4 elements. It is not a
string as the array does not end with a NULL character ('\0'). Lets read the next section to understand
the reason.

5.2.2 STRINGS
In our earlier sections we have read about character strings. Strings are created using single dimension
array of characters (char/wchar_t). The NULL character (ASCII value 0) marks the end of the string.
Each string is represented using a single one-dimensional array. The array-size of the character array
determines the number of characters the string can hold including the terminating NULL character. So,
the character array,
ANSI: char name[11];
Unicode: wchar_t name[11];

is capable of holding 11 characters. If 'name' is to be used as a string, the last character needs to be a
NULL character, which leaves only 10 elements starting from the beginning to be used for storing other
characters. So, if we want to store the string “C LANGUAGE”, it will be represented in the array as below:

63
Quick & Indepth C With Data Structures

Each character is stored in a separate array element with the last element having the '\0' (NULL)
character. So, for strings we must leave one position for the NULL character and thus, calculate the
array-size accordingly. Now, let us consider the following six character array declaration and
initialization:
1. char rgchLang[] = { 'C', ' ', 'L', 'A', 'N', 'G', 'U', 'A', 'G', 'E' };
2. char rgchLang[10] = { 'C', ' ', 'L', 'A', 'N', 'G', 'U', 'A', 'G', 'E' };
3. char rgchLang[11] = { 'C', ' ', 'L', 'A', 'N', 'G', 'U', 'A', 'G', 'E' };
4. char rgchLang[11] = { 'C', ' ', 'L', 'A', 'N', 'G', 'U', 'A', 'G', 'E', '\0' };
5. char rgchLang[] = “C LANGUAGE”;
6. char rgchLang[11] = “C LANGUAGE”;

The first and second statements declare the array 'rgchLang' to be of 10 elements. Both are not strings
as both do not end with a NULL character. The third statement declares the array to be of size 11
elements, but has only the first 10 elements filled with the initialization values. As we have discussed in
the previous section, the last element of the array is automatically initialized to 0 (NULL) because we
have not specified any initialization value for that element. So, the array declared in the third statement
is a string. The fourth statement too declares a string as we have specifically terminated the array with
a NULL character during the initialization. The fifth and sixth statements also declare strings. When we
initialize character arrays within double quotes as in the fifth and sixth statements, the compiler
automatically terminates the array with a NULL character. In the fifth statement, the compiler calculates
the required array-size to be of the initialization string's length plus a terminating NULL character. This
makes it 11 in our example.

5.3 TWO-DIMENSIONAL ARRAYS

We may have situations where we need to store a table of values. If we need to store the marks of all the
subjects of an entire class, a one-dimensional array will not serve the purpose. We will need a two-
dimensional array which will help us represent the marks as a table, with each student in a row and
mark in each individual subject in a column. A two dimensional array has the following general form:
Data-Type ArrayName[Total-Rows][Total-Columns];

The array elements are accessed as 'ArrayName[Row][Column]'. While storing values, we generally
store each row at a time i.e. one row is completed before moving to the next row. Similar to one-
dimensional arrays, the row and column subscripts start from 0 (zero) and continue till 'Total-Rows – 1'
for rows and 'Total-Columns – 1' for columns.

It is worth noting that C follows a row-major ordering for representing multi-dimensional arrays (two-
dimensional and higher) unlike some languages like 'Fortran' which follow a column-major ordering.
Row-major order and column-major order are methods for storing multidimensional arrays in linear
storage such as random access memory. In a row-major order, the consecutive elements of a row reside
next to each other (like [0][0], [0][1], [0][2], [1][0], [1][1], [1][2], [2][0], [2][1], [2][2], ...). Similarly in
a column-major order, the consecutive elements of a column reside next to each other (like [0][0], [1]
[0], [2][0], [0][1], [1][1], [2][1], [0][2], [1][2], [2][2], ...).

64
 ARRAYS

A two-dimensional array is logically represented as given in the next diagram.

Though it is worth knowing that in actual memory the two-dimensional array is represented in a linear
contiguous order and not as a table as we see it. What we see in the above image is the logical
representation and not the physical representation. In physical representation, array elements occur
sequentially with all columns of each row appearing together. So, all columns of row 0 will come first
followed by the columns of row 1 and so on. We need to be bothered about the logical representation
only, as C internally converts the logical representation to physical representation and vice versa. If we
are accessing the array elements using a pointer (points to the memory location of the data), we may
need to work as per the physical representation. We will study about pointers in our later chapters.

For our above example, we may have a table similar as below (Total Marks in each subject:100).
Mathematics Physics Chemistry Biology
Student 1 78 85 80 71
Student 2 92 89 70 68
Student 3 67 72 84 81
Student 4 75 70 73 79
Student 5 71 67 64 68
Student 6 94 91 82 76
Student 7 83 77 67 71
Student 8 76 78 85 82
Student 9 88 73 83 78
Student 10 67 71 68 70

Now, let us write a program which accepts the above information from the user and stores them in an
array. At the end, the program print the marks and the aggregate percentage obtained by a student.

65
Quick & Indepth C With Data Structures

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define TOTALSUBJECTS 4
#define TOTALSTUDENTS 10

int main(void)
{
/* Variable Declaration & Initialization */
int iCounterStudents = 0;
int iCounterSubjects = 0;
int iAggregate = 0;
int rgiMarks[TOTALSTUDENTS][TOTALSUBJECTS] = { 0 };
const char rgszSubjects[TOTALSUBJECTS][12] = { "Mathematics",
"Physics",
"Chemistry",
"Biology" };

/* Accept the marks obtained by all the students. */

while (iCounterStudents < TOTALSTUDENTS)
{
printf("Student #%02d\n===========\n", iCounterStudents + 1);
iCounterSubjects = 0;

while (iCounterSubjects < TOTALSUBJECTS)

{
printf("Enter the marks for %s: ", rgszSubjects[iCounterSubjects]);
scanf("%d", &rgiMarks[iCounterStudents][iCounterSubjects]);
iCounterSubjects++;
}

printf("\n");
iCounterStudents++;
}

iCounterStudents = 0;
iCounterSubjects = 0;

/* Now, display the marks and the percentage obtained by all the students. */
/* We will first print the headings and then the marks. */
printf("\n\t\t\t\tSTUDENT MARKS\n\t\t\t\t*************\n\t\t");
while (iCounterSubjects < TOTALSUBJECTS)
{
/* Make each subject column fixed width (11 chars) & left aligned. A tab is also
inserted between two columns to provide proper separation between columns. */
printf("%-11s\t", rgszSubjects[iCounterSubjects++]);
}

/* After the subject headings, print the heading for the percentage column. */
printf("%%");

/* Now, print the marks and the percentage obtained by the students. */
while (iCounterStudents < TOTALSTUDENTS)
{
printf("\nStudent #%02d:\t", iCounterStudents + 1);
iCounterSubjects = 0;
iAggregate = 0;

while (iCounterSubjects < TOTALSUBJECTS)

{
printf("%02d \t", rgiMarks[iCounterStudents][iCounterSubjects]);

66
 ARRAYS

/* Calculate the aggregate of all the subjects for the student. */

iAggregate += rgiMarks[iCounterStudents][iCounterSubjects];

/* Move to the next subject marks for the student. */

iCounterSubjects++;
}

/* Print the percentage obtained by this student. */

printf("%02.02lf", (double) iAggregate / (double) TOTALSUBJECTS);
/* Move to the next student. */
iCounterStudents++;
}

printf("\n\n");

return 0;
}

The program will generate the following output:

STUDENT MARKS
*************
Mathematics Physics Chemistry Biology %
Student #01: 78 85 80 71 78.50
Student #02: 92 89 70 68 79.75
Student #03: 67 72 84 81 76.00
Student #04: 75 70 73 79 74.25
Student #05: 71 67 64 68 67.50
Student #06: 94 91 82 76 85.75
Student #07: 83 77 67 71 74.50
Student #08: 76 78 85 82 80.25
Student #09: 88 73 83 78 80.50
Student #10: 67 71 68 70 69.00

Now, let us understand what we are doing in the above program.

We have declared two counters 'iCounterStudents' and 'iCounterSubjects'. 'iCounterSubjects' is
incremented to help us move to the next subject when accepting or printing the marks. Similarly,
'iCounterStudents' is incremented to help us move to the next student. We store the marks in the
'iCounterSubjects' subject column for each student row. The marks obtained by each student are stored
in the 'iCounterStudents' row corresponding to the respective student.

We use a while loop (iteration statements) for accepting or printing the values. We will study about
loops in our later chapters. For a brief understanding, loops help us execute a group of statements
(stated within the loop block) repeatedly. 'while loops' execute the group of statements enclosed within
the while block (marked by the opening and ending curly braces just following the while statement)
repeatedly till the while condition is true. The while condition is the expression following the while
statement within the parentheses. In our program, the expressions like 'iCounterStudents <
TOTALSTUDENTS' and 'iCounterSubjects < TOTALSUBJECTS' are the while conditions. An example of the
while loop block in the above program is:
while (iCounterSubjects < TOTALSUBJECTS)
{
printf("%-11s\t", rgszSubjects[iCounterSubjects++]);
}

67
Quick & Indepth C With Data Structures

For scenarios where we need to both increment a counter and also check for the loop condition, 'for
loops' (explained in later chapters) are better. But, for simplicity of understanding we have used the
while loops over here.

In the first while loop, we are accepting the marks obtained by all the students of the class. We have
used two while loops over here with one inside the other. The outer while block enumerates through all
the students of the class. The inner while block accepts the marks obtained by a student in all the
subjects. So in simple terms, the outer while block loops through all the students of the class and the
inner while block loops through all the subjects for each student.

In the next while block we are printing the subject names as headings. Note, that we are looping
through the string array 'rgszSubjects' to print the subject names. A string array is actually a two
dimensional character array with each row representing a string, and each column representing a single
character of that string. We have declared the array as 'const' as we do not expect the array's contents
to change. We are printing the subject names in fixed length size of 11 characters and as left-aligned.
This makes our output format remain consistent. We have separated each heading by a 'tab' ('\t')
character. Please refer to section 2.4.3 and 4.3 for special characters and format specifiers respectively.

In the final while block we are printing the marks obtained by the students. Over here too, we are using
two while blocks – one within another. As discussed earlier, the outer while block enumerates through
all the students and the inner while block enumerates through all the subjects for a student. The inner
while block also calculates the aggregate of all the marks obtained by each student, and the outer while
block prints the aggregate percentage obtained by the student.

5.3.1 INITIALIZATION OF TWO-DIMENSIONAL ARRAYS

Like a one-dimensional array, a two-dimensional array may be initialized with a list of values (enclosed
in braces) during its declaration. The general form of initialization of a two-dimensional array is:
Data-Type ArrayName[Total-Rows][Total-Columns] = { { comma separated list of values for Row 0 },
{ comma separated list of values for Row 1 },
.
.
.
{ comma separated list of values for Row N } };

Where the comma separated values are assigned to the column elements of a row sequentially in the
provided order.
int rgiNumbers[3][3] = { { 62, 81, 23 },
{ 34, 98, 23 } };

In the above example, the value 62 is assigned to 'rgiNumbers[0][0]', 81 is assigned to 'rgiNumbers[0]

[1]', 23 is assigned to 'rgiNumbers[0][2]', 34 is assigned to 'rgiNumbers[1][0]', 98 is assigned to
'rgiNumbers[1][1]' and 23 is assigned to 'rgiNumbers[1][2]'. The other three elements 'rgiNumbers[2]
[0]', rgiNumbers[2][1]' and 'rgiNumbers[2][2]' are automatically assigned 0 (zero) by the compiler.

68
 ARRAYS

When we are initializing the array during declaration, the total row count becomes optional. If omitted,
the total row count becomes equal to the number of rows specified during the initialization. So, if we
change our above example as below.
int rgiNumbers[][3] = { { 62, 81, 23 },
{ 34, 98, 23 } };

The array will now have 2 rows and 3 columns.

5.3.2 STRING ARRAYS

We have read about strings in section 5.2.2. An array of strings is called a string array. As each string is
represented in a one-dimensional array, an array of such strings is represented using a two-dimensional
array. Each string within the array is stored in each row of the array. A string array is nothing but a two-
dimensional character (char/wchar_t) array and is declared as:
ANSI string array: char Array-Name[Total-Rows][Total-Columns];
Unicode string array: wchar_t Array-Name[Total-Rows][Total-Columns];

The 'Total-Columns' must be greater than or equal to the largest character string (including the
terminating NULL character) within the entire array. Just like any other two-dimensional array, we can
also initialize the string array during declaration.
char Names[5][11] = { “John”,
“Somshubhro”,
“Mary” };

In the above declaration cum initialization, the first 3 rows of the array will have the strings “John”,
“Somshubhro” and “Mary”. The remaining 2 rows will have 0 (zero) length (blank - “”) strings.
If we are initializing the array during declaration, we may skip specifying the total row size of the array.
In such a case the total row count becomes equal to the number of rows specified during the
initialization. If we change our above example to:
char Names[][11] = { “John”,
“Somshubhro”,
“Mary” };

We will now have 3 rows in the array.

5.4 MULTI-DIMENSIONAL ARRAYS

C allows arrays of more than two dimensions. The exact limit depends on the compiler. The general form
of a multi-dimensional array is:
Data-Type ArrayName[Size 1][Size 2][Size 3]....[Size N];
where 'Size n' is the size of the nth dimension. Some examples of multi-dimensional arrays are:
int rgiValues[10][7][5];
const float rgfltRates[5][3][2];

69
Quick & Indepth C With Data Structures

Remember: The sizeof operator we had discussed in section 3.9 can also be used on arrays. When the
sizeof operator is passed an array, it returns the total memory (in bytes) used by the array. Consider the
following integer array:
int rgiValues[10][7][5];
The array will have '10 * 7 * 5' elements and will occupy '10 * 7 * 5 * sizeof(int)' bytes of memory.
Assuming ‘int’ occupies 4 bytes, the sizeof operator will return '10 * 7 * 5 * 4’ as the total memory size
consumed by the array.

Within our program, we can calculate the number of elements of an array using the sizeof operator as:
‘sizeof(rgiValues)/sizeof(int)’. The expression will return the total number of elements in all dimensions
of the array (10 * 7 * 5).

The expression ‘sizeof(rgiValues)/sizeof(rgiValues[0])’ will return the array size in the first dimension of
the array i.e. 10 in our case.

We can define a macro which will return the number of elements in a single dimension array as:
#define ARRAYSIZE(x) sizeof(x)/sizeof(x[0])

We can use the macro as below:

int rgiMarks[30];
int iNumElements = ARRAYSIZE(rgiMarks); /* Will return 30 */

5.5 SAMPLE PROGRAMS

1. Accept numbers from the user and sort the given numbers in ascending order.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_NUMBERS 10

int main(void)
{
/* Variable Declaration & Initialization */
int rgiNumbers[MAX_NUMBERS] = { 0 };
int iCounter1 = 0;
int iCounter2 = 0;
int iMinimum = 0;
int iSwap = 0;

/* Accept the numbers from the user. */

while (iCounter1 < MAX_NUMBERS)
{
printf("Enter the number: ");
scanf("%d", &rgiNumbers[iCounter1]);
iCounter1++;
}

70
 ARRAYS

/* Now, sort the numbers. We go till second last element. We do not iterate till the last
element because before we start iterating the second last element, all elements till that
element are already sorted and have values less than or equal to the second last and the
last elements. Only the second last and last elements need sorting between themselves
which is done during the second last iteration itself i.e. if the last element is found
to have a value less than the second last element, their values are swapped. */
iCounter1 = 0;
while (iCounter1 < MAX_NUMBERS – 1)
{
/* We start assuming that we have no elements with value less than this element. If
we find any, we will mark that as minimum. */
iMinimum = iCounter1;
iCounter2 = iCounter1 + 1;

/* All elements before 'iCounter1' are already sorted and have values less than
or equal to the elements which are yet to be sorted. Now, try to find the
next minimum value within the remaining unsorted elements.*/
while (iCounter2 < MAX_NUMBERS)
{
if (rgiNumbers[iCounter2] < rgiNumbers[iMinimum])
{
/* We found a new minimum value. */
iMinimum = iCounter2;
}

/* Increase the inner counter. */

iCounter2++;
}

/* Swap the 'iCounter1' element's content with the element with the least value
within the unsorted elements. If 'iCounter1' and 'iMinimum' are the same, we
will have no swapping and it will remain the same. */
iSwap = rgiNumbers[iMinimum];
rgiNumbers[iMinimum] = rgiNumbers[iCounter1];
rgiNumbers[iCounter1] = iSwap;

/* Increase the outer counter. */

iCounter1++;
}

/* Print the sorted sequence. */

printf("\n\nSorted list: ");
iCounter1 = 0;
while (iCounter1 < MAX_NUMBERS)
{
printf("%d%s", rgiNumbers[iCounter1++],
((MAX_NUMBERS - 1) == iCounter1) ? "\n\n" : ", ");
}

return 0;
}

We are doing selection sort in our program. The selection sort algorithm sorts an array by
repeatedly finding the minimum element (considering ascending order) from the unsorted part
of the array and adding it to the sorted part. The algorithm maintains two imaginary sub-arrays
within a given array.
• The sub-array which is already sorted at the left of the array.
• Remaining sub-array in the right which is unsorted.

71
Quick & Indepth C With Data Structures

In every iteration of selection sort, the minimum element (considering ascending order) from the
unsorted sub-array is picked and swapped with the initial element in the unsorted sub-array.
Initially, the sorted sub-array is empty and the unsorted sub-array is the entire input list. The
algorithm proceeds by finding the smallest (or largest, depending on sorting order) element in
the unsorted sub-array, swapping it with the leftmost unsorted element say 'A'. We then
continue the process from the element following this element 'A' i.e. from 'A + 1'. Assume we
have the array as 5, 7, 8, 2. In our first iteration, the entire array is unsorted. The minimum
value '2' within the array is found and swapped with '5' resulting in the array becoming 2, 7, 8,
5. Now in our next iteration, we start from '7' and swap it with the next minimum value '5'
resulting in the array becoming 2, 5, 8, 7. Now, start from '8' and find the next minimum value
i.e. '7'. We swap '7' and '8' resulting in the array becoming 2, 5, 7, 8, which is the sorted array.

2. Accept a string from the user and reverse the string in-place i.e. within the same array.
#include <stdio.h>

#define MAX_STRING 100

int main(void)
{
/* Variable declaration and initialization. */
char szName[MAX_STRING + 1] = { 0 };
char chSwap = 0;
int iCounter1 = 0;
int iCounter2 = 0;

/* Accept the string from the user. */

printf("Enter a string within %d characters: ", MAX_STRING);
scanf("%s", szName);

/* Make 'iCounter2' move to the end of the string. */

while (szName[iCounter2] != '\0')
{
iCounter2++;
}

/* Decrement 'iCounter2' to make it point to the last NON NULL character. */

iCounter2 = (0 < iCounter2) ? (iCounter2 - 1) : 0;

/* Now, reverse the string. */

while (iCounter1 < iCounter2)
{
/* Swap the element values. */
chSwap = szName[iCounter1];
szName[iCounter1] = szName[iCounter2];
szName[iCounter2] = chSwap;

/* Update the counters. */

iCounter1++;
iCounter2--;
}

/* Print the reversed string. */

printf("\nThe reversed string is: %s.\n\n", szName);

return 0;
}

72
 ARRAYS

For reversing the string we need to start from both ends of the string/array. We will move
'iCounter1' from the beginning of the array and 'iCounter2' from the end. We will keep swapping
the 'iCounter1' and 'iCounter2' element values. We will move 'iCounter1' forward and
'iCounter2' backwards after each swap. We will keep moving till 'iCounter1' is less than
'iCounter2', because by the time they cross, we have already reversed the entire string.
Assuming our string to be “ABCDE”, 'iCounter1' will initially point to 'A' and 'iCounter2' will point
to 'E'. When they are swapped our string becomes “EBCDA”. Now, 'iCounter1' is incremented
and 'iCounter2' is decremented, making 'iCounter1' point to 'B' and 'iCounter2' point to 'D'.
After swapping, the string now becomes “EDCBA”. Again, 'iCounter1' is incremented and
'iCounter2' is decremented. Their values become equal and we stop. We get the reversed string
as 'EDCBA'.

For making 'iCounter2' point to the end of the string, we need to increment 'iCounter2' till it
finds the terminating NULL ('\0') character. 'iCounter2' will then be decremented by 1 to point
to the last NON NULL character in the string.

3. Accept 10 numbers from the user and calculate the 'mean' and 'standard deviation' of the
numbers. 'Mean' means the average of the numbers. 'standard deviation' is a measure that is
used to quantify the amount of variation or dispersion of a set of data values. A low standard
deviation indicates that the values tend to be close to the mean of the set, while a high standard
deviation indicates that the values are spread out over a wider range. For a finite set of
numbers, the standard deviation is found by taking the square root of the average of the
squared deviations of the values from their average value. The 'deviation' is obtained by
squaring the difference between the value and the mean. The average of these deviations is
obtained to get the 'variance'. The square root of the variance is the standard deviation.
/* Header Files */
#include <stdio.h>
#include <math.h>

/* Defined Values */
#define MAX_NUMBERS 10

int main(void)
{
/* Variable declaration and initialization. */
int rgiNumbers[MAX_NUMBERS] = { 0 };
int iCounter = 0;
double dblMean = 0;
double dblTotalDeviation = 0;
double dblStandardDeviation = 0;

/* Accept the numbers from the user. */

while (iCounter < MAX_NUMBERS)
{
printf("Enter the number: ");
scanf("%d", &rgiNumbers[iCounter]);
iCounter++;
}

73
Quick & Indepth C With Data Structures

/* Calculate the mean. We increment the counter in a single statement

as the addition. As we have done a post-increment, the counter will be
incremented after the addition is performed. */
iCounter = 0;
while (iCounter < MAX_NUMBERS)
{
dblMean = dblMean + rgiNumbers[iCounter++];
}
dblMean = dblMean / MAX_NUMBERS;

/* Calculate the total deviation. */

iCounter = 0;
while (iCounter < MAX_NUMBERS)
{
dblTotalDeviation += pow(rgiNumbers[iCounter] - dblMean, 2);
iCounter++;
}

/* Calculate the standard deviation. */

dblStandardDeviation = sqrt(dblTotalDeviation / MAX_NUMBERS);

/* Print the results. */

printf("\nThe mean is %.06lf and the standard deviation is %.06lf.\n",
dblMean, dblStandardDeviation);

return 0;
}

We have used two math functions 'pow' and 'sqrt' in our program. These functions are builtin
functions and are declared in the header file 'math.h'. So, we have included the header file
'math.h' in our program.
The 'pow' function finds the power of a provided base value. It takes two arguments, where the
first argument is the base value whose power is to be found and the second argument is the
exponent of the power. It returns the power as a 'double'. It has the following declaration:
double pow (double base, double exponent);
The 'sqrt' function finds the square root of the provided argument value. It returns the square
root as a 'double'. It has the following declaration:
double sqrt (double argValue);

4. Write a program which accepts numbers from the user and stores them in a 3-dimensional
array. It then prints the array and the sum of all the entered numbers.
#include <stdio.h>

/* Defined Values */
#define SIZEX 2
#define SIZEY 3
#define SIZEZ 4

int main(void)
{
/* Variable declaration and initialization. */
int rgiNumbers[SIZEX][SIZEY][SIZEZ] = { 0 };
int i = 0;
int j = 0;
int k = 0;
long lSum = 0;

74
 ARRAYS

/* Accept the numbers from the user. */

while (i < SIZEX)
{
j = 0;
while (j < SIZEY)
{
k = 0;
while (k < SIZEZ)
{
printf("Enter rgiNumbers[%d][%d][%d]: ", i, j, k);
scanf("%d", &rgiNumbers[i][j][k]);
k++;
}

j++;
}

i++;
}

/* Print the numbers and calculate the sum. */

i = 0;
while (i < SIZEX)
{
j = 0;
while (j < SIZEY)
{
k = 0;
while (k < SIZEZ)
{
printf("\nrgiNumbers[%d][%d][%d]: %d", i, j, k, rgiNumbers[i][j][k]);
lSum = lSum + rgiNumbers[i][j][k];
k++;
}

j++;
}

i++;
}

/* Now, print the sum. */

printf("\nThe sum of all the numbers is: %ld.\n\n", lSum);

return 0;
}

There will be a total of SIZEX x SIZEY x SIZEZ (2 x 3 x 4 = 24) numbers accepted from the user.

75
Quick & Indepth C With Data Structures

CONDITIONAL STATEMENTS AND BRANCHING

[CHAPTER-6]

6.1 INTRODUCTION
Programs do not always execute statements sequentially in the order in which they appear. Situations
may arise where we may have to change the order of execution of statements depending on certain
conditions. We can incorporate decision making ability depending on conditions within our C programs.
These kind of statements are called decision control statements and they are:
• if, if-else, if-else ladder and nested if statements
• switch statement
• Conditional or Ternary operator statement
• goto statement

6.2 DECISION CONTROL USING if STATEMENT

'if' statement is followed by an expression which evaluates a condition, and if the result of the condition
check is true (non-zero), it performs the operations stated within the 'if' block. The 'if' block is the
group of statements specified between the opening and closing curly brace pair just following the 'if'
statement. 'if' statements take the following form:
if (test condition)
{
Statements to execute if condition check is true.
}

The 'if' statement can be visualized as the decision block (diamond) within a flowchart.

76
 CONDITIONAL STATEMENTS AND BRANCHING

In most computer Languages including C, 'true' means a non-zero value and 'false' means a value of
zero.

If the test-condition expression requires a relational operator but none has been specified, then by
default the expression is compared for inequality with 0 (zero). Consider the following examples:
if (x)
{
printf(“x is not equal to 0.”);
}

if (1)
{
printf(“This statement will always be printed.”);
}

if (0)
{
printf(“This statement will never get printed.”);
}

There can be multiple statements within the 'if' block. If the test-condition is true, the statements within
the 'if' block will be executed and if the test-condition is false, they will be skipped and the execution
will jump to the statements following the 'if' block. When the test-condition is true, both the statements
within the 'if' block and the statements following the 'if' block are executed in sequence.

If the number of statements within the 'if' block is limited to one, then the curly brace pair bounding the
'if' block becomes optional. So, we may have something like this:
if (x < 0)
printf(“The value of x is negative.”);

Let us write a C program which accepts the rate and quantity to be purchased of an item and calculates
the total amount to be paid by the customer. It also gives a discount of 5% to a regular customer.
/* Header Files */
#include <stdio.h>
#include <ctype.h>

/* Defined Values */
#define DISCOUNT_RATE 5

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uQuantity = 0;
double dblRate = 0;
double dblAmount = 0;
char chDiscount = 0;

77
Quick & Indepth C With Data Structures

printf("Enter the quantity of the item to be purchased: ");

scanf("%u", &uQuantity);

printf("\nEnter the rate of the item to be purchased: ");

scanf("%lf", &dblRate);

while (toupper(chDiscount) != 'Y' && toupper(chDiscount) != 'N')

{
printf("\nApply discount (y/n): ");
chDiscount = getchar();
}

dblAmount = uQuantity * dblRate;

if (toupper(chDiscount) == 'Y')
{
/* Apply discount. */
dblAmount = dblAmount * ((double)(100 - DISCOUNT_RATE) / (double)100);
}

printf("\n\nThe total amount due from the customer is %.02lf.\n\n", dblAmount);

return 0;
}

In the above program we have used a C library function 'toupper'. The function accepts a character and
returns its uppercase value. If the provided character is already uppercase, the same is returned. It has
the following form:
int toupper (int c);

6.3 DECISION CONTROL USING if-else STATEMENT

The if-else statement extends the 'if' statement for providing the ability to perform a set of operations
when the 'if' condition is not met. It has the following form:
if (test condition)
{
Statements to execute if the condition check is true.
}
else
{
Statements to execute if the condition check is false.
}

If the result of the test-condition check is 'true', the statements within the 'if' block are executed and if
it is 'false', the statements within the 'else' block are executed. So, either the if-block or the else-block is
executed but not both. The if-else control can be visualized as the next flowchart:

78
 CONDITIONAL STATEMENTS AND BRANCHING

If the number of statements within the 'if' block or the 'else' block is limited to one, then the curly brace
pair bounding the block becomes optional. But, it is recommended that we use curly braces as it
improves readability and is considered a good programming practice. So, we may have:
if (x < 0)
printf(“The value of x is negative.”);
else
printf(“The value of x is positive.”);

Let us write a C program which checks whether a given number is odd or even.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber = 0;

printf("Enter the number: ");

scanf("%u", &uNumber);

if (uNumber % 2 == 0)
{
/* The number is divisible by 2 i.e. it leaves a
remainder of 0 when divided by 2. So, it is even. */
printf("\nThe number is even.\n\n");
}
else
{
/* The number is odd. */
printf("\nThe number is odd.\n\n");
}

return 0;
}

79
Quick & Indepth C With Data Structures

6.4 DECISION CONTROL USING if-else LADDER STATEMENTS

We may have situations where we need to check multiple conditions and perform operations depending
on the condition which comes true. The 'if-else' ladder statements allow us to do just that. The test
conditions are checked one after the other till a condition comes true, upon which the control moves
within the condition block which has come true. It has the following form:
if (test condition 1)
{
Statements to execute if the condition check 1 is true.
}
else if (test condition 2)
{
Statements to execute if the condition check 2 is true.
}
else if (test condition 3)
{
Statements to execute if the condition check 3 is true.
}
…….
…….
…….
…….
else if (test condition N)
{
Statements to execute if the condition check N is true.
}
else
{
Statements to execute if all the above condition checks are false.
}

The 'if-else' ladder can be visualized as the next flowchart:

80
 CONDITIONAL STATEMENTS AND BRANCHING

The 'test condition 1' is checked for validity and if 'false', the program checks 'test condition 2'. If that
too is 'false', the program checks 'test condition 3'. This continues till a condition is found to be 'true'.
When a satisfied condition is found, the control moves within the block corresponding to the condition,
and all statements corresponding to that block are executed. Thereafter, all test condition checks
following the satisfied condition are skipped and control moves to the statements following the 'if-else'
ladder. If the 'test condition 1' is found to be 'true', the statements corresponding to 'test condition 1'
are executed and then the control jumps to the statement following the 'if-else' ladder, skipping rest of
the condition checks. If none of the test condition checks satisfy, the control moves directly within the
'else' block of the 'if-else' ladder. If we do not have any 'else' block, the control jumps to the statement
following the 'if-else' ladder.

Let us write a C program which accepts the sides of a triangle from the user. It first checks whether the
provided dimensions can form a triangle. If they form a triangle, it checks whether the triangle is an
equilateral triangle or an isosceles triangle or a scalene.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
double dblSide1 = 0;
double dblSide2 = 0;
double dblSide3 = 0;

81
Quick & Indepth C With Data Structures

/* Accept the dimensions from the user. */

printf("Enter the length of the first side: ");
scanf("%lf", &dblSide1);

printf("Enter the length of the second side: ");

scanf("%lf", &dblSide2);

printf("Enter the length of the third side: ");

scanf("%lf", &dblSide3);

if (dblSide1 <= 0 || dblSide2 <= 0 || dblSide3 <= 0)

{
/* In a triangle no side can be of zero length. */
printf("\nThe provided dimensions do not form a triangle.\n\n");
}
else if ((dblSide1 + dblSide2) <= dblSide3 ||
(dblSide1 + dblSide3) <= dblSide2 ||
(dblSide2 + dblSide3) <= dblSide1)
{
/* In a triangle the addition of the length of any
two sides is always greater than the third. */
printf("\nThe provided dimensions do not form a triangle.\n\n");
}
else if (dblSide1 == dblSide2 && dblSide2 == dblSide3)
{
/* A triangle is equilateral, if all its sides are equal in length. */
printf("\nThe triangle is an equilateral triangle.\n\n");
}
else if (dblSide1 == dblSide2 || dblSide1 == dblSide3 || dblSide2 == dblSide3)
{
/* A triangle is isosceles, if any of its two sides are equal. */
printf("\nThe triangle is an isosceles triangle.\n\n");
}
else
{
/* A triangle with all unequal sides is called scalene. */
printf("\nThe triangle is scalene.\n\n");
}

return 0;
}

6.5 DECISION CONTROL USING NESTED if STATEMENTS

What if we need to make more decisions within an 'if' or 'else' block itself. In such cases we can use
more 'if' control statements within an existing 'if' block. As a matter of fact, we may use any kind of valid
C control statements within an 'if' or 'else' block including other 'if-else' statements. When 'if-else'
statements are nested within outer 'if-else' statements, we call them nested 'if' statements.
if (test condition 1)
{
if (test condition 2)
{
Statements;
}
else
{

82
 CONDITIONAL STATEMENTS AND BRANCHING

Statements;
}
}
else
{
if (test condition 3)
{
Statements;
}
else
{
Statements;
}
}

Let us write a C program which accepts three numbers from the user and finds the greatest of the three
numbers.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
int iValue1 = 0;
int iValue2 = 0;
int iValue3 = 0;

printf("Enter the first value: ");

scanf("%d", &iValue1);

printf("Enter the second value: ");

scanf("%d", &iValue2);

printf("Enter the third value: ");

scanf("%d", &iValue3);

if (iValue1 >= iValue2)

{
if (iValue1 >= iValue3)
{
printf("\nThe largest number is %d.\n\n", iValue1);
}
else
{
printf("\nThe largest number is %d.\n\n", iValue3);
}
}
else
{
if (iValue2 >= iValue3)
{
printf("\nThe largest number is %d.\n\n", iValue2);
}

83
Quick & Indepth C With Data Structures

else
{
printf("\nThe largest number is %d.\n\n", iValue3);
}
}

return 0;
}

6.6 DECISION CONTROL USING switch STATEMENT

When we need to make decisions depending on many possible values of a variable or expression, we
can use 'if-else' ladders for the purpose. But if the possible alternatives are many, we will end up with a
long list of statements, making our program not only complex but unreadable. For such situations, C
provides the 'switch' statement. The 'switch' statement tests the value of a variable or expression
against a list of 'case' values. When a match is found, a block of statements corresponding to the
matching case is executed. The general form of a 'switch' statement is given below:
switch (expression)
{
case value1:
Statements to be executed when the result of expression is value1.
break;
case value2:
Statements to be executed when the result of expression is value2.
break;
….
….
default:
Statements to be executed when the result of expression does not match
any of the above values.
break;
}

The result of the 'expression' must be an integer or a character value. The 'case' values 'value1',
'value2', ... are constants or constant expressions (resulting in an integer result) and are called 'case
labels'. Each of these values must be unique within the 'switch' statement. There can be 0 (zero) or
more statements corresponding to a case label. There is no need to include braces around these
statements, though it is a good practice to use braces when there are multiple statements. Every case
label must end with a colon ( : ).

When the 'switch' block is encountered, the result of the expression is compared sequentially against
the 'case' values. If a match is found, the block of statements corresponding to the matching case label
is executed. The 'break' statement at the end of each case label block causes an exit from the 'switch'
statement and the control jumps to the statement following the 'switch'. If the 'break' statement is not
provided or missed, the control will thereupon move to the next case label, and the block corresponding
to that case label will also get executed. The control does not break out of the 'switch' until a 'break'

84
 CONDITIONAL STATEMENTS AND BRANCHING

statement is found or the end of the 'switch' statement is encountered.

The 'default' is an optional label. It is executed when the value of the expression does not match any of
the case values. If not present, no action takes places and the control moves to the statement following
the 'switch' block.

Now it is time for us to understand the 'switch' statement with the help of a program.

A local footwear shop has a reward system for all its customers depending on the amount of purchase
made by them during a year. The customers are divided into six categories ranging from category 0 to
5. Category 1 to 5 are for repeat customers who made purchases ranging from $100 to $1000 during
the year. First time customers or customers who made purchases of less than $100 are placed under
category-0. All customers are provided a certain percentage of discount depending on the category to
which the customer belongs. The discount is calculated on the marked price and are as follows:
Category Discount Percentage
Category-0 2-4
Category-1 7
Category-2 9
Category-3 11
Category-4 14
Category-5 15

The discount percentage of category-0 customers vary between 2 to 4 depending on the current
purchase amount. A discount of 2% is provided for purchase amount of within $20, 3% for purchases
between $21 to $60 and 4% for purchases above $60. Category-5 customers not only get the discount
percentage assigned to them but also the discount percentage assigned for category-0 customers. So,
for a purchase amount of $80, a category-5 customer will get a discount of (15+4) 19% and for a
purchase amount of $15, the customer will get a discount of (15+2) 17%.

The program accepts the marked price of the total purchase made by a customer and the category of
the customer. It then calculates the net amount payable by the customer after applying the discount.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define CATEGORY_0_DISC1 2
#define CATEGORY_0_DISC2 3
#define CATEGORY_0_DISC3 4
#define CATEGORY_1_DISCOUNT 7
#define CATEGORY_2_DISCOUNT 9
#define CATEGORY_3_DISCOUNT 11
#define CATEGORY_4_DISCOUNT 14
#define CATEGORY_5_DISCOUNT 15

int main(void)
{
/* Variable declaration and initialization. */

85
Quick & Indepth C With Data Structures

double dblMarkedAmount = 0;
double dblNetBillAmount = 0;
unsigned int uDiscount = 0;
unsigned int uCategory = 10;

while (uCategory > 5)

{
/* Keep asking for the category unless the user enters a valid one. */
printf("Enter the customer's category: ");
scanf("%u", &uCategory);
}

printf("Enter the total marked price of the items purchased: ");

scanf("%lf", &dblMarkedAmount);

switch (uCategory)
{
case 1:
uDiscount = CATEGORY_1_DISCOUNT;
break;
case 2:
uDiscount = CATEGORY_2_DISCOUNT;
break;
case 3:
uDiscount = CATEGORY_3_DISCOUNT;
break;
case 4:
uDiscount = CATEGORY_4_DISCOUNT;
break;
case 5:
uDiscount = CATEGORY_5_DISCOUNT;
/* Deliberately falling through as category-5 customers also get the
discount offered to the category-0 customers. So, avoided using the
break statement which makes us fall through to the next case label. */
case 0:
{
if (dblMarkedAmount <= 20)
{
uDiscount = uDiscount + CATEGORY_0_DISC1;
}
else if (dblMarkedAmount <= 60)
{
uDiscount = uDiscount + CATEGORY_0_DISC2;
}
else
{
/* For purchases made above $60. */
uDiscount = uDiscount + CATEGORY_0_DISC3;
}

break;
}
default:
break;
}

dblNetBillAmount = dblMarkedAmount * ((double)(100 - uDiscount) / (double)100);

printf("\nThe total bill amount for the customer is $%.02lf.\n", dblNetBillAmount);

return 0;
}

86
 CONDITIONAL STATEMENTS AND BRANCHING

If the user enters a wrong category, we ask for the category again. We keep doing so until the user
enters a valid category (0-5). As, 'uCategory' is of type 'unsigned int', it can never have a value of less
than 0 (zero). Even if the user enters a value of less than 0 (zero), it is wrapped around and will produce
a value greater than 5.

Note, for the case label 5, we did not use a break statement. This is because category-5 customers also
get the discount available to category-0 customers. So, we have allowed the control to fall through to
the case label 0.

While calculating the net amount, we have typecasted the discount to double, as we need to avoid the
integer round-off.

6.7 DECISION CONTROL USING CONDITIONAL (TERNARY) OPERATOR

We have read about the conditional operator in section 3.8. We will carry on our discussion from there.
We can nest conditional statements where one conditional statement is within another. This can be
explained using a sample program.

Let us write a program which accepts three numbers from the user and prints the greatest number.
/* Header Files */
#include <stdio.h>
int main(void)
{
/* Variable declaration and initialization. */
int iValue1 = 0;
int iValue2 = 0;
int iValue3 = 0;
int iMaxValue = 0;

printf("Enter the first number: ");

scanf("%d", &iValue1);

printf("Enter the second number: ");

scanf("%d", &iValue2);

printf("Enter the third number: ");

scanf("%d", &iValue3);

/* Get the largest number within iMaxValue. */

iMaxValue = (iValue1 >= iValue2)
? (iValue1 >= iValue3) ? iValue1 : iValue3
: (iValue2 >= iValue3) ? iValue2 : iValue3;

printf("\nThe largest number is %d.\n\n", iMaxValue);

return 0;
}

In the program we are using nested conditional operator statements to find the largest of the three
values. '(iValue1 >= iValue2)' is the first (outer) conditional comparison, and if it is 'true' the expression
'(iValue1 >= iValue3) ? iValue1 : iValue3' is evaluated. If it is 'false', the expression '(iValue2 >=

87
Quick & Indepth C With Data Structures

iValue3) ? iValue2 : iValue3' is evaluated.

If the inner conditional comparison '(iValue1 >= iValue3)' is 'true' we get 'iValue1' else 'iValue3'.
If the inner conditional comparison '(iValue2 >= iValue3)' is 'true' we get 'iValue2' else 'iValue3'.

6.8 DECISION CONTROL USING goto STATEMENT

All the previous decision control statements we discussed control the execution flow depending on
certain specified conditions. Like many other languages, C also supports the unconditional jump
statement 'goto'. 'goto' statement allows us to jump from one location to another within the same
function of a program.

The 'goto' requires a label to be defined within the function in which the 'goto' statement is used. The
label identifies the location to which the jump is to be made. The label is any valid variable name and is
followed by a colon. The label is placed immediately before the statement where the control needs to be
transferred upon a goto call. The goto statement can be of the following forms:

Whenever the control reaches the goto statement, the control jumps to the statement immediately
following the label. When the label is placed after a goto statement, all statements appearing between
the goto call and the label will be skipped. This is called forward jump. On the other hand, if the label is
placed before the goto statement, then we may have a loop kind of situation where all the statements
between the label and the goto call are executed repeatedly. This is called backward jump.

Using a goto statement we can make the control jump out of a control block (like if, switch, while, for
statements), but cannot make it to jump into a control block. The goto label is generally used at the end
of a function and goto calls to this label is made to skip operations when further processing within the
function is not desired (such as when an essential condition is not met).

Though in some cases 'goto' comes in real handy but, for most occasions the use of 'goto' should be
avoided. The biggest drawbacks of the 'goto' statement are:
• It reduces the readability of the program as the jump location can be far off from the goto
statement and we need to keep searching for the goto label within the entire function.
• It becomes difficult to trace or track the control flow of the program, making the program logic
difficult to understand.
• It is error prone to code using goto as it alters the sequential flow of logic and can make
unintended statements to be executed. On the other hand, it can also lead to required

88
 CONDITIONAL STATEMENTS AND BRANCHING

statements getting missed from execution. If care is not taken, we may get into an infinite loop
when using backward jump.

'goto' statements should only be used in situations where the use of structured solutions are far more
complex to be implemented.

Let us write a program which determines whether a specified year is a leap year or not. A year is a leap
year if it is divisible by 4 and is not divisible by 100, or it is divisible by 400.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
int iYear = 0;

ACCEPTYEAR:
printf("Enter the year to check (Enter 1582 or less to stop): ");
scanf("%d", &iYear);

if (iYear <= 1582)

{
/* Stop the program. */
goto STOP;
}

if (iYear % 4 == 0 && iYear % 100 != 0)

{
/* The year is divisible by 4 but not 100. So, it is a leap year. */
printf("The year %d is a leap year.\n\n", iYear);
}
else if (iYear % 400 == 0)
{
/* The year is divisible by 400. So, it is a leap year. */
printf("The year %d is a leap year.\n\n", iYear);
}
else
{
printf("The year %d is not a leap year.\n\n", iYear);
}

/* Accept the next year to be checked for a leap year. */

goto ACCEPTYEAR;

STOP:
printf("\nStopping execution...\n\n");

return 0;
}

In the program, we have two goto statements and labels. When the user enters a year 1582 or less, we
stop the program. We have used the statement 'goto STOP' for doing so. When the statement 'goto
ACCEPTYEAR' is hit, the program jumps to the label 'ACCEPTYEAR'. This makes the program ask the user
to enter the next year to check.

89
Quick & Indepth C With Data Structures

6.9 SAMPLE PROGRAMS

1. We are provided with the length of two sides of a triangle and the angle between these two
sides. We need to find the length of the third side. We can keep the following in mind when
writing the program:
◦ The angle between two sides of a triangle must be less than 180 degrees.
◦ Assume ‘a’, ‘b’ and ‘c’ are the three sides of a triangle, with ‘a’ being the unknown side. ‘A’
is the angle (in degrees) between the sides ‘b’ and ‘c’.

On these assumptions, the length of the side ‘a’ is calculated as:

/* Header Files */
#include <stdio.h>
#include <math.h>

/* Defined Values */
#define PI 3.141592

int main(void)
{
/* Variable declaration and initialization. */
double dblSide1 = 0;
double dblSide2 = 0;
double dblSide3 = 0;
double dblAngle = 0;
double dblRadian = 0;
double dblCosine = 0;

printf("Enter the first side of the traingle: ");

scanf("%lf", &dblSide1);

printf("Enter the second side of the triangle: ");

scanf("%lf", &dblSide2);

printf("Enter the angle between the two sides: ");

scanf("%lf", &dblAngle);

if (dblSide1 == 0 || dblSide2 == 0)
{
printf("\nThe length of a side cannot be 0.\n");
goto STOP;
}
else if (dblAngle >= 180)
{
printf("\nThe angle must be less than 180 (degrees).\n");
goto STOP;
}

90
 CONDITIONAL STATEMENTS AND BRANCHING

/* We have accepted the angle in degrees but cos function in C accepts

the angle is radians. So, we need to convert the degrees to radians. */
dblRadian = (dblAngle * PI) / 180;

dblCosine = (dblSide1 * dblSide1) + (dblSide2 * dblSide2) -

(2 * dblSide1 * dblSide2 * cos(dblRadian));
if (dblCosine < 0)
{
/* We take only the positive square root. */
dblCosine = -dblCosine;
}

dblSide3 = sqrt(dblCosine);
printf("\nThe length of the third side is %.02lf.\n", dblSide3);

STOP:
printf("\nStopping program...\n\n");

return 0;
}

In our program we have used two math functions 'cos' and 'sqrt'. So, we have included the
header file 'math.h'. The 'cos' function returns the cosine of an angle. The function accepts the
angle in radian. As the angle specified by the user is in degrees, we needed to convert it to
radian, so that we can pass it to the 'cos' function. The 'cos' function has the following
declaration:
double cos(double)
The 'sqrt' function returns the square root of a number and has the following declaration:
double sqrt(double);

2. In our next program we try to find all the roots of a quadratic equation. A quadratic equation is
of the form:
ax2 + bx + c = 0

Our program will accept the values of ‘a’, ‘b’ and ‘c’ from the user and then find the roots of the
equation using the following formula:

The program will try to find the roots in two steps. In the first step, it will try to find the
discriminant which is:

When we look at the formula to find the roots of the equation, we notice that the discriminant is
the same value which appears under the square root.

In the second step we find the roots of the quadratic equation depending on the value of the
discriminant.
• If the discriminant is positive, there are two real roots of the quadratic equation.
• If the discriminant is 0 (zero), there are two real roots with equal value and are
calculated as ‘-b/2a’.

91
Quick & Indepth C With Data Structures

• If the discriminant is negative, there are two complex roots, as the square root of a
negative number is a complex number. We will not go into the details of complex
numbers over here. As for a quick understanding a complex number is a number that
can be expressed in the form 'a + bi', where ‘a’ and ‘b’ are real numbers and ‘i’ is an
imaginary unit. The square of ‘i’ (imaginary unit) is -1. We have two parts in a complex
number: ‘a’ is the real part and ‘b’ is the imaginary part. So, the roots of the equation in
this case will be

where,

is the modulus or absolute value of the discriminant.

/* Header Files */
#include <stdio.h>
#include <math.h>

int main(void)
{
/* Variable declaration and initialization. */
double a = 0;
double b = 0;
double c = 0;
double dblDiscriminant = 0;
double dblImaginary = 0;
double dblRoot1 = 0;
double dblRoot2 = 0;

/* Accept the numbers from the user. */

ACCEPTA:
printf("Enter the value of a: ");
scanf("%lf", &a);

if (a == 0)
{
/* The value of 'a' cannot be 0. If it is 0, the equation becomes
linear and does not remain quadratic. */

printf("a cannot be 0. Please re-enter.\n");

goto ACCEPTA;
}

printf("Enter the value of b: ");

scanf("%lf", &b);

printf("Enter the value of c: ");

scanf("%lf", &c);

dblDiscriminant = (b * b) - (4 * a * c);

if (dblDiscriminant > 0)
{
dblRoot1 = (-b + sqrt(dblDiscriminant)) / (2 * a);
dblRoot2 = (-b - sqrt(dblDiscriminant)) / (2 * a);

92
 CONDITIONAL STATEMENTS AND BRANCHING

printf("\nThere are two distinct and real roots, and they are “
“%.02lf and %.02lf.\n", dblRoot1, dblRoot2);
}
else if (dblDiscriminant == 0)
{
dblRoot1 = dblRoot2 = -b / (2 * a);

printf("\nThere are two equal and real roots with the value %.02lf.\n", dblRoot1);
}
else
{
/* The discriminant is negative. We need to use the absolute value of
discriminant when finding its square root. As the discriminant is
negative, we are adding a '-' sign in front to make it positive. */
dblRoot1 = dblRoot2 = -b / (2 * a);
dblImaginary = sqrt(-dblDiscriminant) / (2 * a);

printf("\nThere are two distinct and complex roots, and they are "
"%.02lf + %.02lfi and %.02lf - %.02lfi.\n",
dblRoot1, dblImaginary, dblRoot2, dblImaginary);
}

return 0;
}

93
Quick & Indepth C With Data Structures

CONDITIONAL STATEMENTS WITH LOOPS

[CHAPTER-7]

7.1 INTRODUCTION
Loops help us execute a group of statements repeatedly till a condition called the loop condition is met.
In our previous chapter we have read the use of goto statements to simulate a loop. But, it is very
impractical and inconvenient to implement a loop using goto. Moreover, it is prone to errors due to the
inherent disadvantages of the goto statement.

C language provides us with three kinds of loop constructs and they are:
• The while loop
• The do while loop
• The for loop

Every program loop consists of two components, one is the body of the loop consisting of the group of
statements which needs to be executed repeatedly, and the other being the control condition. The
control condition is a test condition which determines how long the loop is repeated. The loop is
executed repeatedly till the test condition is true. Depending on the location of the control condition in
the loop, a loop construct can be classified as entry-controlled loop or exit-controlled loop. In the entry-
controlled loop, the control conditions are tested before the start of the loop. If the conditions are
satisfied the loop is executed, else it is skipped. In the exit-controlled loop, the control conditions are
tested at the end of the body of the loop. So, the body of the loop is executed at-least once and the first
execution is unconditional. The flowcharts corresponding to the entry-controlled loop and the exit-
controlled loops are given below.

94
 CONDITIONAL STATEMENTS WITH LOOPS

When working with loops we should be careful regarding the control condition. The control condition
should achieve the exact number of loop executions as the logic desires. If the loop runs less or more
than the desired number, we may get invalid results and even a program crash (in case working with
pointers or array indices). Also if the control condition is never reached, we will get into an infinite loop
resulting in the body of the loop getting executed infinitely. So, when working with a loop we should
always do a dry run (with a pen and paper) using some dummy data before doing the actual execution.
The dummy data helps us understand the desired number of loop repetitions that will yield the required
result. It also helps us determine the correct control condition that will provide that number of
repetitions. As a simple example, if we want a loop to be executed 5 times and we start our counter 'i'
from 0 (zero), should we execute the control condition till 'i < 5' or 'i <= 5' ? The correct answer is 'i <
5' , which can easily be evaluated using a dry run. So when writing a loop, we need to take care of the
following:
1. Declaring the counter of the correct data type and initializing it to a proper value.
2. Determine the number of times the body of the loop needs to be executed.
3. Provide the correct control condition which achieves the #2 point.
4. Write the body of the loop which gives us the desired result.
5. Increment the counter.

7.2 THE while LOOP

We have used the while loop in our previous chapters. It is the simplest of all the loop constructs in C.
The while loop is an entry-controlled loop and has the following form:
while (test condition)
{
body of the loop
}

Before the loop is entered, the test condition is evaluated and if true, the body of the loop gets
executed. After the body of the loop is executed, the test condition is again evaluated and if true, the
body is executed again. This continues until the test condition becomes false and the control moves to
the statement following the loop.

The body of the loop may contain any number of statements. But for program readability, it is desirable
to limit the number of statements to the number of lines visible at a time (does not require scrolling) in
the editor or IDE (Integrated Development Environment). We can do so by moving groups of statements
to separate user-defined functions. We will learn about user-defined functions in our later chapters.

The braces of the while loop are required only when we have more than one statement in the loop body.
In case we have only one statement, the braces become optional. Though it is considered a good
programming practice to use the braces on all occasions for better program readability.

In our first example, we try to reverse the digits of a given number. For example, if the entered number
is 12345, the program will produce the reverse as 54321.

95
Quick & Indepth C With Data Structures

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber = 0;
unsigned int uReverse = 0;

/* Accept a number from the user. */

printf("Enter a positive number: ");
scanf("%u", &uNumber);

/* Reverse the number. */

while (uNumber > 0)
{
/* Advance the place value of all digits in uReverse by one place
and put the next digit from uNumber at the unit's place. */
uReverse = uReverse * 10 + uNumber % 10;
uNumber = uNumber / 10;
}

/* Print the result. */

printf("\nThe reverse of the number is %u.\n\n", uReverse);

return 0;
}

We keep iterating till 'uNumber' is greater than 0 (zero), which means we have reversed all the digits
and nothing is left to be reversed.

If the test-condition expression requires a relational operator but none has been specified, then by
default the expression is compared for inequality with 0 (zero). In our above example, we could have
also achieved the same result if we had used the statement 'while (uNumber)' which would have been
equivalent to using 'while (uNumber != 0)'.

7.3 THE do while LOOP

The while loop is an entry-controlled loop as it checks the test condition before executing the loop body.
The body of the loop may not get executed even once, if the test condition is not met at the very first
attempt.

We may need situations where we need to execute the body of the loop at-least once irrespective of
whether the condition is met or not. Such situations demand the use of an exit-controlled loop. For the
exit-controlled conditions, C language provides the loop construct 'do while'. It has the following form:
do
{
body of the loop
} while (test condition);

Unlike other loop constructs, the 'do while' construct has a semi-colon at the end of the test condition.

96
 CONDITIONAL STATEMENTS WITH LOOPS

The program executes the body of the loop and then proceeds to test the control condition. If true, it
executes the body of the loop again. This continues until the test condition becomes false, upon which
the control goes to the statement following the loop block.

Now let us write a program that accepts numbers from the user and calculates their sum. It keeps
accepting the numbers until the user enters 0 (zero) as the number. Even if the user does not want to
run the program, we need to accept the number at-least once when the user gets the chance to enter 0
(zero) to stop the program. This desires the use of an exit-controlled loop.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
int iNumber = 0;
long lSum = 0;

/* We have to accept atleast one number from the user. If it is 0 (zero) we stop. */
do
{
/* Accept the number. */
printf("Enter a number: ");
scanf("%d", &iNumber);

lSum += iNumber;

} while (0 != iNumber);

/* Print the result. */

printf("\nThe sum of all the numbers is %ld.\n\n", lSum);

return 0;
}

7.4 THE for LOOP

Similar to the while loop, the for loop is also an entry-controlled loop but provides a more concise and
convenient loop structure. It has the following form:
for (initialization ; test condition ; increment or decrement)
{
body of the loop
}

The for loop consists of the following components:

• The initialization of the control variables are done in the initialization section of the statement.
The initialization section may optionally contain initialization of other variables used within the
for loop too, but is mostly used for the control variables. Initialization is done by assignment of
initial (starting) values to the variables using assignment statements (example: iCount=idx=0).
We can initialize multiple variables, with each initialization being separated from the other by a
comma. As an example while counting the sum of all numbers starting from 5 (and continuing

97
Quick & Indepth C With Data Structures

till 10) we could be initializing the variables as: iNumber = 5, lSum = 0. The initialization section
is executed only once and is done before any other portion of the loop gets executed.
• After the initialization section, the test condition is evaluated and if true, the body of the loop
gets executed. Unlike the initialization section, the test condition is evaluated at the start of
each iteration, and this continues until the test condition becomes false. The body of the loop
keeps getting executed again and again until the test condition becomes false, upon which the
control goes to the statement following the for loop.
• After the test condition is evaluated, the body of the loop gets executed.
• After executing the body of the loop, the control comes to the 'increment or decrement' section
of the for statement. Here, the control variable is incremented or decremented depending on
the program logic (example: i++ or idx = idx – 1 or a = a + 2). We can increment or decrement
multiple variables, with each increment or decrement being separated from the other by a
comma (example: i++, j--).

Let us write a small C program which uses two variables. It initializes the first variable to 1 and the
second to 100. It prints the value of the first variable from 1 to 100 and the second variable from 100
to 1. We loop till the first variable is less than 101 and the second variable is greater than 0. If any of
the condition is false, we break out of the loop.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
int iNumber1 = 0;
int iNumber2 = 0;

/* iNumber1 starts from 1 and is incremented till it is less than or equal to 100.
So, it prints the number from 1 to 100 and as soon as it crosses 100, the test
condition becomes false. iNumber2 starts from 100 and is decremented till it is
greater than 0. So, it prints the number from 100 to 1, and as soon as it
reaches 0, the test condition becomes false. */
for (iNumber1 = 1, iNumber2 = 100; iNumber1 <= 100 && iNumber2 > 0; iNumber1++, iNumber2--)
{
printf("iNumber1: %d, iNumber2: %d.\n", iNumber1, iNumber2);
}

return 0;
}

The loop in our program gets executed 100 times, printing the number 'iNumber1' from 1 to 100 and
the number 'iNumber2' from 100 to 1. The variable 'iNumber1' is initialized to 1 and the variable
'iNumber2' is initialized to 100 before starting the loop. The test condition 'iNumber1 <= 100 &&
iNumber2 > 0' is evaluated before moving to the loop body. As both the conditions in the test condition
are joined by the logical AND (&&) operator, both need to be true for the entire test condition to be true.
After the body of the loop is executed, the increment or decrement section 'iNumber1++, iNumber2--' of
the for statement gets executed. Thereupon the test condition is evaluated again, and if true the body of
the loop gets executed once more. This continues until the test condition evaluates to false.

98
 CONDITIONAL STATEMENTS WITH LOOPS

As the for loop is an entry-controlled loop, the body of the loop may never get executed if the test
condition is evaluated to false at the very beginning of the loop itself.

In our next program, we will try to find all the factors of a given number. The factors will include 1 and
the number itself. There will be no factors (except the number itself) which will be greater than the half
of the given number. So, we iterate till we are less than or equal to the half of the number 'uNumber / 2'.
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber = 0;
unsigned int uReverse = 0;

/* Accept a number from the user. */

printf("Enter a positive number: ");
scanf("%u", &uNumber);

/* Find factors. */
printf("\nThe factors of the number are:\n");
for (unsigned int i = 1; i <= uNumber / 2; i++)
{
if (0 == uNumber % I)
{
printf("%u, ", i);
}
}

/* Now print the number entered by the user as that is also a factor. */
printf("%u\n\n", uNumber);

return 0;
}

The for statement has the following features:

1. As discussed earlier, the initialization section may initialize multiple variables separated by
commas (example: iNumber = 1, lSum = 0). If both the variables need to be initialized to the
same value we could do something like 'lSum = iNumber = 0'.
2. As we had discussed, the increment or decrement section may also contain more than one
expression that are separated by commas (example: iNumber1++, iNumber2 = iNumber2 – 2).
3. Just like other loop constructs, the test condition may contain multiple conditions joined
together using the logical operators.
4. If any section within the for statement is not required, it may be omitted. However, the semi-
colons separating the sections must remain. If the test condition is omitted, we may get into an
infinite loop, until broken using a break or a goto statement. We will read about break a little
later in this chapter. We may have something like below:
int i = 0;
for (; i < 50; )
{
printf(“The value of i is %d.\n”, i++);
}

99
Quick & Indepth C With Data Structures

5. If the body of the loop has only one statement, we may omit the braces of the for loop. Consider
the next example:
for (i = 0; i < 50; i++)
printf(“The value of i is %d.\n”, i );
6. We may have a for statement without any body of the loop. The statement may just increment a
variable until it reaches a specific value. These are useful when working with arrays, such as
situations where we need to search for an element with a specific value within the array.
Consider the below example where we have an integer array 'rgiValues' of 100 elements. The
array is filled with values and we need to find the first element with a value of 0.
for (idx = 0; idx < 100 && rgiValues[idx] != 0; idx++);
We come out of the loop when 'idx' reaches 100 (means we did not find any element with a
value of 0) or we find an element with the value 0. Such statements without any body of the
loop need to be ended with a semi-colon.

7.5 NESTED LOOPS

What if we need to use another loop inside an outer loop. In such cases we can use more loop control
statements within an existing loop block. As a matter of fact, we may use any kind of valid C control
statements within a loop block including other loop statements. When loop statements are nested within
outer loop statements, we call them nested loop statements. They take the following form:
while (test condition)
{
while (test condition)
{
--------
--------
for (initialization; test condition; increment or decrement)
{
--------
--------
}
--------
--------
}

do
{
---------
---------
} while(test condition);
---------
---------
}

100
 CONDITIONAL STATEMENTS WITH LOOPS

Nested loops are useful in situations where we need to do further iterations within an existing loop.
Accepting or printing values in multi-dimensional arrays are examples in which we will require nested
loop constructs, where each loop will work with a particular dimension of the array. The outermost loop
will work with the first dimension, the first inner loop will work with the second dimension, the second
inner loop will work with the third dimension and so on.
for (i = 0; i < first dimension length; i++)
{
for (j = 0; j < second dimension length; j++)
{
for (k = 0; k < third dimension length; k++)
{
---------
---------
---------
}
}
}

Let us write a program which accepts the element values of two matrices from the user. It then prints
the sum of these two matrices.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MATRIX_ROWS 3
#define MATRIX_COLS 4

int main(void)
{
/* Variable declaration and initialization. */
int rgiMatrix1[MATRIX_ROWS][MATRIX_COLS] = { 0 };
int rgiMatrix2[MATRIX_ROWS][MATRIX_COLS] = { 0 };
int iRow = 0;
int iCol = 0;

/* Accept the first matrix from the user. */

printf("Enter the elements of the first matrix:\n");
for (iRow = 0; iRow < MATRIX_ROWS; iRow++)
{
for (iCol = 0; iCol < MATRIX_COLS; iCol++)
{
printf("Element [%02d][%02d]: ", iRow, iCol);
scanf("%d", &rgiMatrix1[iRow][iCol]);
}
}

/* Accept the second matrix from the user. */

printf("\nEnter the elements of the second matrix:\n");
for (iRow = 0; iRow < MATRIX_ROWS; iRow++)
{
for (iCol = 0; iCol < MATRIX_COLS; iCol++)
{

101
Quick & Indepth C With Data Structures

printf("Element [%02d][%02d]: ", iRow, iCol);

scanf("%d", &rgiMatrix2[iRow][iCol]);
}
}

/* Add the two matrices and print the result. */

printf("\nThe sum of the two matrices is:");
for (iRow = 0; iRow < MATRIX_ROWS; iRow++)
{
printf("\n");
for (iCol = 0; iCol < MATRIX_COLS; iCol++)
{
printf("%9d", rgiMatrix1[iRow][iCol] + rgiMatrix2[iRow][iCol]);
}
}

printf("\n\n");

return 0;
}

7.6 JUMPING WITHIN LOOPS

The group of statements within a loop is repeated until the test condition becomes false. When
executing a loop, it sometimes becomes necessary to break out of the loop or skip a part of the loop
when certain condition occurs. Consider an example where we are searching for a particular value
within an array of 50 elements. We need to loop though the entire array in search of the element with
the given value. This will require us to loop through all the 50 elements of the array. But, we need to
terminate the search as soon as we find the desired value. In such a case we will require to break out of
the loop when the value is found.

7.6.1 THE break STATEMENT

The break statement is a C keyword and when used within a loop, the control comes out of the
innermost loop containing the break statement. We have seen the use of break in switch statement
which makes the control break out of the innermost switch block. Similarly, when the break statement is
used within a loop (while, do while or for), the control breaks out of the innermost loop and goes to the
statement immediately following the loop. In case of nested loops, when the break statement is
encountered in an inner loop, the control comes out of the loop which directly encloses the break
statement.

102
 CONDITIONAL STATEMENTS WITH LOOPS

We can achieve the same functionality of break statement using the goto statement. For that we need
to define a label outside the loop and use the goto statement to jump to that label. When using nested
loops we can use a single goto statement to jump out of multiple loops, unlike the break statement
which breaks out of the innermost loop only. But, it is recommended that we avoid using the goto
statement as a means to break out of a loop in situations where the break statement can be used. The
break statement is not prone to the disadvantages inherent to the goto statement (drawbacks
mentioned in section 6.8). Some example usage of goto statement to break out of loops are:

103
Quick & Indepth C With Data Structures

In our example we accept an array of integers from the user. We search the array for the first
occurrence of a negative number within the array. If found, we print its location.
/* Header Files */
#include <stdio.h>

/* Defined Values */
#define ARRAY_MAX_ELEMENT 20

int main(void)
{
/* Variable declaration and initialization. */
int rgiNumbers[ARRAY_MAX_ELEMENT] = { 0 };
int iCounter = 0;

/* Accept the array from the user. */

for (; iCounter < ARRAY_MAX_ELEMENT; iCounter++)
{
printf("Enter array element [%02d]: ", iCounter);
scanf("%d", &rgiNumbers[iCounter]);
}

/* Search for the first negative number within the array. */

for (iCounter = 0; iCounter < ARRAY_MAX_ELEMENT; iCounter++)
{
if (rgiNumbers[iCounter] < 0)
{
/* We found it. No need to continue any more. */
break;
}
}

if (iCounter == ARRAY_MAX_ELEMENT)
{
/* We reached the end of the array. */

104
 CONDITIONAL STATEMENTS WITH LOOPS

printf("There are no negative numbers within the array.\n\n");

}
else
{
printf("The first negative number is at the index location %d.\n\n", iCounter);
}

return 0;
}

7.6.2 THE continue STATEMENT

Just like the break statement, continue is also a C keyword and is used to skip a portion of the loop.
But, unlike the break statement which makes the control break out of the loop, the continue statement
makes the control jump to the next iteration by skipping any statements in-between. Whenever it is
encountered, the control directly jumps to the beginning of the loop for the next iteration, skipping the
statements following the continue statement in the current iteration. For the for loop, continue
statement causes the increment and test condition of the loop to execute. For the while and do while
loops, continue statement causes the control to pass to the test condition.

105
Quick & Indepth C With Data Structures

We can achieve the same functionality of continue statement using goto. For that we need to define a
label inside the loop and just after the loop statement. Consider the following example:
while (…....…)
{
LOOP:
----------
----------
if (test condition)
{
goto LOOP;
}
---------
}

In our next program we will generate a pyramid of stars. It will accept the number of rows of the
pyramid from the user subject to a maximum rows of 30. The program will keep printing the pyramids
until the user enters 0 (zero) as the number of rows to print. It will print a similar pattern for each user
input:
*
***
*****
*******
*********
***********

To keep symmetry, the number of stars in each row has to be two more than the previous row, with each
extra star extending on both sides of the previous row. So if the top of the pyramid has one star, we will

106
 CONDITIONAL STATEMENTS WITH LOOPS

have odd number of stars in every row of the pyramid (1, 3, 5, ...). The number of stars in each row is
one less than the twice of the row number i.e. if the row number is 'a', the number of stars in this row
will be '2a - 1'. For each row, spaces will be inserted at the beginning to help the stars align to the
middle of the pyramid.
/* Header Files */
#include <stdio.h>

/* Defined values */
#define MAX_ROWS 30

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uRows = 0;
unsigned int uCurRow = 0;
unsigned int idx = 0;
unsigned int uSpaces = 0;
unsigned int uStars = 0;

do
{
/* Accept the number of rows to print. */
printf("Enter the number of rows of the pyramid: ");
scanf("%u", &uRows);

if (0 == uRows)
{
break;
}
else if (uRows > MAX_ROWS)
{
printf("Please enter a row count within %d.\n\n", MAX_ROWS);
continue;
}

printf("\nThe pyramid of stars:\n");

/* Print each row at every iteration. */
for (uCurRow = 1; uCurRow <= uRows; uCurRow++)
{
/* Print the preceding spaces so as the stars appear
at the middle of the next row to come. */
uSpaces = uRows – uCurRow;
for (idx = 1; idx <= uSpaces; idx++)
printf(" ");

/* Now print the stars. */

uStars = 2 * uCurRow – 1;
for (idx = 1; idx <= uStars; idx++)
printf("*");

/* Move to the next row. */

printf("\n");
}

printf("\n\n");

} while (0 != uRows);

return 0;
}

107
Quick & Indepth C With Data Structures

7.7 CONCISE TEST CONDITIONS

Just like if conditions, in loops too, if the test-condition expression requires a relational operator but
none has been specified, then by default the expression is compared for inequality with 0 (zero). We
need not make explicit comparisons with 0 (zero). Consider the following:

while ( expression != 0 )
is equivalent to using
while ( expression )

AND

while ( expression == 0 )
is equivalent to using
while ( !expression )

Example:
while ( x != 0 && y % 2 == 0) is same as
while ( x && ! (y % 2))

for (x = 10; x != 0; x--) is same as

for (x = 10; x; x--)

7.8 SAMPLE PROGRAMS

1. Let us write a C program which will show the multiplication tables of numbers. The program will
accept two numbers from the user. Multiplication tables of all numbers starting from the first
given number to the second will be printed. The program uses nested for loops where the first
for loop iterates through the numbers whose multiplication tables are shown. The second for
loop shows the multiplication table corresponding to each number.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber = 0;
unsigned int uEndNumber = 0;
unsigned int uMultiply = 0;

/* Accept the number from which to start showing the multiplication tables.
The number must be between 1 and 99. If less than 1 or greater than 99,
we ask the user to again enter the number. */
while (uNumber < 1 || uNumber > 99)
{
printf("The number from which to start the multiplication tables (1-99): ");

108
 CONDITIONAL STATEMENTS WITH LOOPS

scanf("%u", &uNumber);
}

/* Accept the number up to which the multiplication tables need to be shown.

The number must be between 1 and 99. If less than 1 or greater than 99,
we ask the user to again enter the number. The number must also be greater
than or equal to the number from which to start. */
while (uEndNumber < 1 || uEndNumber > 99 || uEndNumber < uNumber)
{
printf("The last number till which multiplication tables to show (1-99): ");
scanf("%u", &uEndNumber);
}

/* Loop through all the numbers and show their multiplication table. Each
multiplication table is shown in each row with each product being shown
separated from the other by a comma. To properly show the results, we
are left-aligning our output and using a fixed length 4-character output
with space as the padding character. */
printf("\nThe multiplication table of the numbers are:");
for (; uNumber <= uEndNumber; uNumber++)
{
printf("\n%2u: ", uNumber);

/* Show the multiplication table of uNumber. */

for (uMultiply = 1; uMultiply <= 10; uMultiply++)
{
if (uMultiply < 10)
{
printf("%4u, ", uNumber * uMultiply);
}
else
{
printf("%4u", uNumber * uMultiply);
}
}
}

printf("\n\n");

return 0;
}

2. In our next program we will accept a number from the user and determine whether the number
is prime or composite. We will keep accepting numbers from the user until the user enters 0
(zero) as the number.

To check whether a number is prime or not, we need to check for the number's divisibility till
the square root of the number. If a number n is not a prime (i.e. a composite), it can be factored
into two factors 'A' and 'B' as below:
N = A * B;
where 'A' and 'B' are whole numbers. If both 'A' and 'B' were greater than the square root of 'N',
then 'A * B' would have been greater than 'N'. So, both 'A' and 'B' together cannot be greater
than 'N'. If say the number 'A' is greater than the square root of 'N', then 'B' must be less than
the square root of 'N'. So, if we check for divisibility of 'N' till the square root of 'N', then we
must have also checked its divisibility by 'B'. If it is divisible, 'N' is not a prime and is a
composite. So, checking its divisibility by 'A' becomes a non-requirement.

109
Quick & Indepth C With Data Structures

/* Header Files */
#include <stdio.h>
#include <math.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber = 0;
unsigned int uMaxFactor = 0;
unsigned int uFactor = 0;
unsigned char IsPrime = 0;

do
{
/* Accept the number which needs to be checked for prime. */
printf("Enter the number to check for prime: ");
scanf("%u", &uNumber);

if (0 == uNumber)
{
/* The user does not want to continue. */
break;
}
else if (1 == uNumber)
{
printf("The number '1' is neither a prime nor composite.\n\n");
continue;
}

uMaxFactor = (unsigned int) sqrt((double) uNumber);

for (IsPrime = 1, uFactor = 2; uFactor <= uMaxFactor; uFactor++)
{
if (0 == (uNumber % uFactor))
{
/* The number is divisible by uFactor. So, it is not a prime.
Break out of this inner loop. */
IsPrime = 0;
break;
}
}

if (IsPrime) /* It is the same as 'if (IsPrime != 0)'. */

{
printf("The number '%u' is a prime number.\n\n", uNumber);
}
else
{
printf("The number '%u' is a composite number.\n\n", uNumber);
}

} while (0 != uNumber);

printf("\n\n");

return 0;
}

3. Let us write a program which accepts two positive numbers from the user and finds the HCF and
LCM of the two numbers. To find the HCF of the two numbers, we use the Euclidean algorithm.
Here, we continuously replace the larger of the two numbers by its remainder when divided by

110
 CONDITIONAL STATEMENTS WITH LOOPS

the smaller of the two numbers. This is repeated until the remainder comes as 0 (zero), upon
which the last divisor (for which the remainder came as 0 [zero]) is the HCF. The LCM is
calculated as the product of the two numbers divided by their HCF.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber1 = 0;
unsigned int uNumber2 = 0;
unsigned int uModulo1 = 0;
unsigned int uModulo2 = 0;
unsigned int uHCF = 0;
unsigned long uLCM = 0;

do
{
/* Accept the first number. */
printf("Enter the first number: ");
scanf("%u", &uNumber1);

} while (0 == uNumber1);

do
{
/* Accept the second number. */
printf("Enter the second number: ");
scanf("%u", &uNumber2);

} while (0 == uNumber2);

/* Find the HCF using the Euclidean algorithm. */

uModulo1 = uNumber1;
uModulo2 = uNumber2;
while (uModulo1 != 0 && uModulo2 != 0)
{
if (uModulo1 > uModulo2)
{
uModulo1 = uModulo1 % uModulo2;
}
else
{
uModulo2 = uModulo2 % uModulo1;
}
}

if (0 == uModulo1)
{
uHCF = uModulo2;
}
else
{
uHCF = uModulo1;
}

/* Calculate the LCM. The LCM is calculated as the product of the

two numbers divided by their HCF. */
uLCM = (uNumber1 * uNumber2) / uHCF;

/* Print the results. */

111
Quick & Indepth C With Data Structures

printf("\nThe HCF of the numbers is %u and LCM is %lu.\n\n", uHCF, uLCM);

return 0;
}

4. Let us write a program which accepts a positive number and prints its binary representation. To
do this, we need to loop through each bit of the number from left to right, determine if the bit is
0 or 1, and print its value. We keep doing so until we have printed all the bits.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber = 0;
int idx = 0;
unsigned int uBit = 0;

/* Accept the number from the user. */

printf("Enter the number: ");
scanf("%u", &uNumber);

printf("The binary equivalent of the number is:\n");

for (idx = (sizeof(uNumber) * 8) - 1; idx >= 0; idx--)
{
/* We are evaluating each bit of the number one by one starting from
the most significant bit (MSB) (i.e. the left-most bit) and
continuing towards the right. At every iteration, we are doing a
right-shift of the current bit in such a manner so that it becomes
the least significant bit (LSB) (i.e. the right-most bit).
|MSB|0|1|0|1|1|0|......|LSB|.
We are then doing a bitwise AND of it with 1 which gives 1 if the
LSB is 1, else gives 0. So, in effect we are just trying to find
whether the current bit is 1 or 0 and printing it accordingly.
Remember: The value of uNumber does not change due to the shift or
bitwise AND, as it is not the LValue (i.e. it does not appear at
the left of the assignment operator). So, its original value is
retained and does not change as nothing gets assigned to it.
*/
uBit = (uNumber >> idx) & 1;
printf("%u", uBit);
}

printf("\n\n");

return 0;
}

5. In the next example we will accept a number from the user and find its factorial.
/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
unsigned short uNumber = 0;
unsigned short uCounter = 0;
unsigned long uFactorial = 0;

112
 CONDITIONAL STATEMENTS WITH LOOPS

/* Accept the number from the user. */

printf("Enter the number: ");
scanf("%hu", &uNumber);

/* Generate the factorial. */

for (uFactorial = 1, uCounter = uNumber; uCounter > 1; uCounter--)
{
uFactorial *= uCounter;
}

/* Print the result. */

printf("\nThe factorial of %hu is %u.\n\n", uNumber, uFactorial);

return 0;
}

6. Let us write a program which prints the Fibonacci series.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_NUM_COUNT 47

int main(void)
{
/* Variable declaration and initialization. */
unsigned int uNumber1 = 0;
unsigned int uNumber2 = 1;
unsigned int uTemp = 0;
unsigned short uCount = 0;

ACCEPTCOUNT:
/* Accept the count of the Fibonacci numbers to print. */
printf("Enter the count of Fibonacci numbers: ");
scanf("%hu", &uCount);

if (uCount < 1 || uCount > MAX_NUM_COUNT)

{
/* At count 47, we get the largest number that can be held
in an 'unsigned int' variable, after which we will start
witnessing an overflow and the value will roll over. */
printf("Please enter a count between 1-47\n\n");
goto ACCEPTCOUNT;
}

/* Print the numbers. */

printf("\nThe Fibonacci numbers are:\n");
for (; uCount > 0; uCount--)
{
printf("%u, ", uNumber2);
uTemp = uNumber2;
uNumber2 += uNumber1;
uNumber1 = uTemp;
}

printf("\n\n");

return 0;
}

113
Quick & Indepth C With Data Structures

POINTERS

[CHAPTER-8]

8.1 INTRODUCTION
C is one of the very few languages which allow direct access to memory locations. With this C brings
one of the powerful features of assembly languages right into a high-level language. In other words C
provides the convenience of a high-level language with the power of an assembly language. We access
the memory locations of variables and functions using pointers. Pointers can be very confusing in the
beginning but, is a very powerful feature which comes really handy. To be able to code well in C, we
must have a very good understanding of pointers.

The entire computer memory is divided into memory cells where each cell is capable of holding 1 byte
of data. Each such memory cell is assigned an address by the operating system. This address helps the
operating system know where to store the data or from where to fetch the data. Each such memory cell
has a numeric address and the maximum size of this address depends on the computer architecture and
the operating system. Many a times we hear about 32-bit/64-bit processor and operating system. In a
32-bit (4 bytes) system, the address is a numeric value which can be stored in 4 bytes (a maximum
address value of 4,29,49,67,295 or 0xFFFFFFFF). So, if the first memory location has an address of 1,
the last one will have an address of 4,29,49,67,295 (0xFFFFFFFF). Similarly in a 64-bit system, the
address is stored in 8 bytes (a maximum address value of 1,84,46,74,40,73,70,95,51,615 or
0xFFFFFFFFFFFFFFFF).

Every variable in a program is assigned a memory location in which the value of the variable is stored.
The size of the memory space (i.e. the number of consecutive memory cells the variable occupies)
depends on the data type of the variable. The address of the variable refers to the starting address of
the memory location which is assigned for the variable. Consider the following example:
int iValue = 978562;
Assuming int takes 4-bytes, the data will be stored in 4 consecutive memory cells with each cell
capable of holding 1 byte of data. Let us assume that the system has chosen the address 100 as the
starting location for this variable. On these assumptions, the variable may be stored as shown in the
following diagram (the order of the bytes depend on the computer architecture):

The first memory byte holds the value 130, the second byte holds 238, the third holds 14 and the fourth

114
 POINTERS

holds 0. The four bytes collectively hold 978562 as a single value for the entire variable. If the variable
is assigned a starting memory location of 100 the four bytes belonging to the variable will have
addresses of 100, 101, 102 and 103 respectively. During execution, the system assigns the address
100 corresponding to the variable name ‘iValue’. We may access the value using its variable name
‘iValue’ or its address ‘100’. Since the memory addresses are just numbers, they can be assigned to
variables and these variables can be used for later direct access of the memory locations. Say in our
above example, we may have a variable called ‘ptr’ which is assigned the memory location of the
variable ‘iValue’ which in our case is 100. So, ‘ptr’ will have the value 100 assigned to it. Later on, we
may use this variable ‘ptr’ to directly access and manipulate the value contained in ‘iValue’. Variables
like ‘ptr’ are called pointers. In other words, pointers are variables that contain the memory addresses
of other variables or functions within our program.

Pointers can be used to directly access the memory locations and read or write values to that location.
As a pointer points to the memory location of a variable or function, thus their name. Since pointers are
also variables, they too occupy a memory location and have an address of their own. As pointers have
memory addresses, we may declare pointers which point to other pointers as well (pointer to pointer).
We will come to this a little later in this chapter. For now let us understand this concept using a simple
diagram.

We have the variable ‘iValue’ which holds the value 978562 and is located at the (starting) memory
address 100. The pointer ‘ptr’ is another variable which has the value 100 (the memory address of
iValue) and is itself located at the memory address 656.

Some of the advantages of using pointers are:

• They increase the execution speed of the program as we have direct access to the memory
locations.
• A pointer helps us to access and manipulate variables which are defined outside the function.
• Pointers are more efficient and convenient to use when dealing with arrays and character
strings.
• Pointers help us to allocate memory dynamically during execution. This helps us to allocate
memory when needed and thus, reduces memory wastage. In simple words, it reduces the
memory needed for a program to run. We will learn about dynamic memory allocation in our
later chapters.
• It is almost impossible to work with some data structures like linked lists, trees etc. without the
help of pointers.
• In many cases, pointers reduce the length and complexity of a program.

115
Quick & Indepth C With Data Structures

Some of the disadvantages of using pointers are:

• If pointers are not used carefully, we may end up accessing and manipulating wrong memory
locations leading to program crashes.
• C does not check the validity of a pointer location and allows us free use of the location. Pointer
arithmetic involves calculation of locations using pointers. If we are wrong in our calculations, C
will not stop us and we may overrun or under run our memory location leading to incorrect
results.
• If dynamically allocated memory is not freed properly, it may result in memory leak.
• Pointer syntax is complex and a little difficult to master, especially pointer to pointer. We may
generate wrong results if we use wrong pointer syntax.

8.2 DECLARING, INITIALIZING AND ASSIGNING POINTERS

Like every other variable in C, pointers too belong to specific data types. Pointer variables need to be
declared as pointers and their data type needs to be specified during declaration. This enables C to
know that these variables are going to be used as pointers and the kind of data they will be pointing to.
The data type of a pointer enables the compiler to determine the operations that will be allowed on the
pointer variable. The declaration of a pointer variable takes the following form:
Data-Type *pointer_name;
The above statement tells the compiler three things:
• The asterisk (*) tells the compiler that this variable (pointer_name) is a pointer.
• pointer_name will point to a variable of type Data-Type.
• Just like any other variable, the pointer_name is assigned a memory location with a memory
address of its own. The size of the memory space assigned to the pointer depends on the
computer architecture. In a 32-bit system, it will use 32 bits (4 bytes) of space and in a 64-bit
system, it will use 64 bits (8 bytes).

Examples:
int *ptr1 → ptr1 is a pointer variable which will point to an integer (int) variable.
float *ptr2 → ptr2 is a pointer variable which will point to a floating point (float) variable.
char *ptr3 → ptr3 is a pointer variable which will point to a character (char) variable.

We may declare a pointer of void data type. This tells the compiler that this pointer is capable of
pointing to a variable of any data type. As the data type of variable pointed by this pointer is
unspecified, many operations on this pointer variable are restricted. For example, we have to typecast
this variable to a specific data type before performing any arithmetic, assignment or increment
operations. void pointers are used in situations where we are not sure regarding the type of data this
variable will point to when we are writing the code. An example declaration of a void pointer is:
void *ptr;

The memory address of a variable is only known during the execution of the program and is not known
to us during writing of the code. So C provides us with an operator (&) which is called the ‘address of’
operator. We have seen it being used in functions like ‘scanf’ in our previous chapters. When the
operator & immediately precedes a variable in a statement, it returns the address of that variable.

116
 POINTERS

Consider the following example:

int iValue = 56;
int *ptr;
ptr = &iValue;

In the above statements, we are declaring a pointer ‘ptr’ of type int and assigning the address of the
integer variable ‘iValue’ to it. The statement ‘ptr = &iValue’ assigns the address of the variable ‘iValue’
to the pointer ‘ptr’. If the address of the variable ‘iValue’ is 100, that address is assigned to the pointer
‘ptr’. The & operator can be used with a simple variable or an array element. Consider the following
example which uses an array element:
int rgValues[] = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};
int *ptr;
ptr = &rgValues[3];

Assuming int takes 4 bytes of memory space, if ‘rgValues[0]’ starts from the address 100, the 4th
element (rgValues[3]) will be located at the memory address 112 (100 + 3 * 4). So in our case 112 will
be assigned to the pointer ‘ptr’.

Some of the invalid usage of the & operator may be:

• int rgValues[] = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};
int *ptr;
ptr = &rgValues;

Array names themselves point to the starting address of the array and thus, they cannot be
used with the & operator. To get the starting address, we may do any of the following valid
operations:
◦ ptr = rgValues;
◦ ptr = &rgValues[0];

• int *ptr;
ptr = &56;

Address of a constant cannot be retrieved using the & operator.

Just like any other variable, pointer variables can also be initialized during their declaration itself.
Pointer variables can also be initialized to NULL (ASCII values 0) signifying that they are currently not
pointing to any memory location. As a matter of fact, we may assign NULL to a pointer variable at any
moment within our program, upon which it signifies that the pointer from thereon does not point to any
memory location. Consider the following example:
int iValue = 475;
int *ptr1 = &iValue;
int *ptr2 = NULL;

117
Quick & Indepth C With Data Structures

if (NULL == ptr2)
{
ptr2 = ptr1;
}

As we can see in the above example, a pointer can also be compared with equality or inequality with
NULL. The first pointer ‘ptr1’ is assigned the address of the variable ‘iValue’ during its declaration itself.
As the ‘iValue’ variable is an integer (int), the pointer ‘ptr1’ is also of integer type. The pointer ‘ptr2’ is
declared and initialized to NULL. Within our if condition, we are checking whether ‘ptr2’ is NULL (does
not point to any valid memory location), and if so, we are assigning the value of ‘ptr1’ to ‘ptr2’. The
assignment makes ‘ptr2’ contain the address of ‘iValue’ and thus, thereupon both ‘ptr1’ and ‘ptr2’ point
to ‘iValue’.
Remember: Before a pointer is initialized or assigned an address, it should not be used as it contains a
garbage value. So, it is always a good practice to initialize a pointer to NULL during its declaration. We
must ensure that the pointer points to the corresponding type of data before performing any operation
using the pointer.

8.3 ACCESSING A VARIABLE USING ITS POINTER

Once a pointer has been assigned the address of a variable, the value of the variable can be accessed
using the pointer. This is done with the help of the indirection operator * (asterisk). Using this operator,
we can not only access the value of a variable (which the pointer is pointing to) but also change it.
Consider the following example:
1. int iValue = 728;
2. int iNumber = 0;
3. int *ptr = NULL;
4. ptr = &iValue;
5. iNumber = *ptr;
6. *ptr = *ptr + 10;
7. (*ptr)++;
8. iNumber++;

In our above example, we are declaring two integer variables and a pointer. Let us find out what
happens in rest of the statements.
• In statement 4, we are assigning the address of the variable ‘iValue’ to the pointer ‘ptr’.
• In statement 5, we are using the indirection operator * to retrieve the integer value contained in
the memory location pointed by the pointer ‘ptr’. As ‘ptr’ is pointing to ‘iValue’, the * operator
returns the value stored in ‘iValue’. This value is assigned to the variable ‘iNumber’. Using ‘*ptr’
is equivalent to using ‘iValue’ itself.
• In statement 6, on the right hand side, we are accessing the value of ‘iValue’ and adding 10 to
it. We are then assigning the value back to the location where ‘ptr’ is pointing to. So in effect,
we are updating ‘iValue’ with this new value. Note, the use of the indirection operator * does not
update the pointer but, the value in the memory location pointed by the pointer. The pointer
keeps pointing to the same memory location as before.
• In statement 7, we are not incrementing or changing the pointer but, the value in the memory
location pointed by the pointer, which in our case is ‘iValue’. So, ‘iValue’ gets incremented after

118
 POINTERS

this statement. But, a statement like *(ptr++) or *ptr++ would have incremented the memory
address pointed by the pointer and tried to access the value contained in that address. In our
case, it may generate a run-time error as that memory address does not belong to us or may be
invalid.
• In statement 8, we are simply incrementing the value of ‘iNumber’. Incrementing this variable
does not have any effect on the pointer or the location pointed by it.

Let us now write a small program which adds two numbers and prints the result. It will do all the
operations using pointers.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
int iNumber1 = 0;
int iNumber2 = 0;
long lSum = 0;
int *pNumber1 = &iNumber1;
int *pNumber2 = &iNumber2;
long *pSum = &lSum;

/* Accept the numbers from the user. */

printf("Enter the first number: ");
scanf("%d", pNumber1);

printf("Enter the second number: ");

scanf("%d", pNumber2);

*pSum = *pNumber1 + *pNumber2;

printf("\nThe sum of the two numbers is %ld.\n\n", *pSum);

return 0;
}

8.4 ARRAY OF POINTERS AND POINTER TO AN ARRAY

8.4.1 ARRAY OF POINTERS

Array of pointers is very similar to declaring array of any other type. Each element of such an array is
capable of pointing to the memory address of a variable which is of the same data type as the pointer.
The general form for declaring an array of pointers is:
Data-Type *array_name[Dimension1-Size][Dimension2-Size]….;
We assign the address of a variable to an element of (a single dimensional) array of pointers in the
following way:
array_name[element-index] = &variable_name;
Let us explain this with a simple example. We assume a single-dimensional array of integer pointers
‘ptr’ with size of 5 elements. So, it is capable of pointing to 5 integer variables. We declare the array and
assign the address of variables ‘iValue1’, ‘iValue2’ and ‘iValue3’ to the first three elements of the array

119
Quick & Indepth C With Data Structures

as below:
int *ptr[5];
ptr[0] = &iValue1;
ptr[1] = &iValue2;
ptr[2] = &iValue3;

Have a look at the below diagram which shows five variables with memory addresses of 100, 112,
116, 6752 and 6756 respectively. The elements of the array point to the addresses of these variables.

We can access the value contained in the location pointed by an element of the array using the
indirection operator * (asterisk) as below:
*array_name[array-index]

8.4.2 POINTER TO AN ARRAY

Pointer to an array is a single pointer variable which is capable of pointing to an entire array. The
declaration of the pointer must include the size of the array to which it will point. The general form of
declaring pointer to a single dimension of an array is:
Data-Type (*pointer_name)[Array-Dimension-Size];
This kind of pointers point to an entire dimension of an array. The array must be of the same data type
as the pointer. If the array is a single-dimensional array, the pointer points to the entire array. In case of
a two-dimensional array, the pointer will point to an entire row of the array. If incremented, the pointer
points to the next row. The general form of assigning the address of an array to such a pointer is:
pointer_name = &array_name; When ‘pointer_name’ has the same number of dimensions as the
array ‘array_name’.
Example [One-dimensional]:
int rgiValues[5] = { 1, 2, 3, 4, 5 };
int (*prgValues)[5] = &rgiValues;
pointer_name = &array_name[0]; When ‘pointer_name’ points to an entire row of a two
dimensional array ‘array_name’ i.e. it points to only a single
dimension of the array.
Example [Pointer points to Row-0 of the array]:
int rgiMatrix[3][4] = { { 1, 2, 3, 4 },
{ 5, 6, 7, 8 },
{ 9, 10, 11, 12 } };
int (*pMatrix)[4] = &rgiMatrix[0];

120
 POINTERS

We can access (read and assign) an element of the array pointed by such a pointer as below:
Read → Variable_Name = (*pointer_name)[column-index];
Assign → (*pointer_name)[column-index] = Value;

Another way to access an element of the array is to add the column index of the element with the
pointer value. This makes it point to the required element. Thereupon, we may use the indirection
operator * (asterisk) to get or set the value of the element.
Memory Address → *pointer_name + column-index
Element Value → *(*pointer_name + column-index)

Let us understand the above using a diagram. Assume we have an integer array ‘rgiValues’ of 8
elements and a pointer ‘ptr’ which points to this array.

Let us write a program in which we are going to use an array of pointers to the elements of an integer
array. We will print the element values using their respective pointers.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define NUM_COUNT 5

int main(void)
{
/* Variable declaration and initialization. */
int idx = 0;
int iNumbers[NUM_COUNT] = { 1, 2, 3, 4, 5 };
int *pNumbers[NUM_COUNT];

/* Assign the address of the elements to the pointers. */

for (idx = 0; idx < NUM_COUNT; idx++)
{
pNumbers[idx] = &iNumbers[idx];
}

/* Now, print the numbers. */

printf("The numbers are: ");
for (idx = 0; idx < NUM_COUNT;)
{
printf("%d", *pNumbers[idx++]);

121
Quick & Indepth C With Data Structures

if (idx < NUM_COUNT)

{
/* Do not print ',' after the last element. */
printf(", ");
}
}
printf("\n\n");

return 0;
}

Let us write another program which utilizes array of pointers. But this time around, these pointers will
point to character strings. String is an array of characters where the last element is a NULL character
(ASCII value 0). The NULL value marks the end of the string. We will declare an array of pointers and
initialize its elements to point to the string constants. We will then print these strings using the
pointers.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_CITIES 10

int main(void)
{
/* Variable declaration and initialization. */
int idx = 0;
const char *pCities[MAX_CITIES] = { "New Delhi",
"Washington, D.C.",
"London",
"Moscow",
"Tokyo",
"Ottawa",
"Canberra",
"Gaborone",
"Nairobi",
"Brasilia"
};

printf("The list of cities are:\n");

for (idx = 0; idx < MAX_CITIES; idx++)
{
printf("%s\n", pCities[idx]);
}

printf("\n");

return 0;
}

8.5 POINTER ARITHMETIC

Pointers can be incremented and decremented just like any other variable. We can also add or subtract
a number from a pointer to directly point to a specific memory location. We can also subtract one
pointer from another giving us the number of items between the two pointer locations.

The result of such pointer arithmetic depends on the data type of the pointer involved in the expression.

122
 POINTERS

When a character pointer is incremented, it will point to the next memory byte address as ‘char’ data
takes one byte space. Whereas when a float pointer is incremented, it will shift by four memory bytes as
‘float’ uses 4 bytes of memory. This comes in real handy when we are working with contiguous memory
locations like an array. We may declare a pointer which points to the elements of an array. The pointer is
incremented or decremented for accessing the array elements, and we need not bother about the
memory size of each element of the array. When incremented, the pointer automatically jumps to the
next element of the array irrespective of the memory space used by each array element.

Remember: If we use the sizeof operator with a pointer argument, the operator returns the size of the
pointer i.e. the total memory used by the pointer itself. It does not return the size of the variable
pointed by the pointer. The size of a pointer depends on the architecture of the program, and will be 4
bytes for a 32-bit program and 8 bytes for a 64-bit program.

We can typecast the memory address of a data type to a pointer of another data type. We need to use
explicit typecast for the purpose. Consider the below example:
unsigned int uValue = 16649;
char *pLetter = (char *) &uValue;
pLetter++;
printf(“\nThe second byte is the letter %c.\n\n”, *pLetter);

In the above example, we are assigning the address of the ‘unsigned int’ variable to a ‘char’ pointer. As
the data type of the pointer is different from the data type of the variable, we have to explicitly typecast
the variable’s address to the pointer’s data type before assigning. After the address has been typecast-
ed to the ‘char’ pointer, incrementing the pointer will make it point to the next memory byte address.
Though we have assigned the address of an ‘unsigned int’ variable (which takes 4 or 8 bytes of memory
depending on the compiler and architecture) to the pointer, yet, incrementing the pointer will shift the
pointer by only 1 byte as the pointer is of ‘char’ data type (which takes 1 byte of memory). In our above
example, the pointer will point to the second byte within the variable which contains the value 65.
Hence, the ‘printf’ statement will print the character ‘A’ which corresponds to the ASCII value 65.

To understand pointer arithmetic better, let us have a look at the following three scenarios.
1. In the first scenario, we have a character array 'rgVals[8]' and two character pointers 'ptr1' and
'ptr2'. 'ptr1' points to the first element of the array and 'ptr2' points to the last element. We
assume that the array starts from the memory location 100. On these assumptions, we evaluate
the result of various pointer arithmetic operations.
2. In the second scenario, we have an integer array 'rgVals[8]' and two integer pointers 'ptr1' and
'ptr2'. 'ptr1' points to the first element of the array and 'ptr2' points to the last element. We
assume that 'int' takes 4 bytes of memory and our array starts from the memory location 100.
3. In our last scenario, we have an array 'rgVals[8]' of double data type and two pointers 'ptr1'
and 'ptr2' of the data type double. 'ptr1' points to the first element of the array and 'ptr2'
points to the last element. We assume that the array starts from the memory location 100. All
results are calculated assuming 'ptr1' is pointing to address 100 and 'ptr2' is pointing to the
address 156.

123
Quick & Indepth C With Data Structures

In the above diagram we can see that addition and subtraction from a pointer moves the pointer
depending on the data type of the pointer. If it is of type ‘char’ it moves by 1 byte and if its ‘double’ it
moves by 8 bytes. We also notice that the name of the array is actually a const pointer to the starting
address of the array and the address of an element of the array can be obtained using the element
index as ‘&array[element index]’.

When we have a pointer pointing to a multidimensional array, we need to keep in mind that in memory a
multidimensional array is represented sequentially. For example, in case of a two-dimensional array all
the columns of row-0 come first, then all the columns of row-1, then row-2 and so on. All the columns
for a row appear together and then the columns of the next row. So in effect the entire two-dimensional
array appears sequentially in the actual physical memory. Assuming we have a two-dimensional array
called ‘rgiValues’ which has 3 rows and 4 columns and is declared as:
int rgiValues[3][4] = { { 7, 14, 2, 6 },
{ 11, 28, 5, 25 },
{ 43, 8, 30, 72 } };

The above array will be represented as the below diagram in actual physical memory.

124
 POINTERS

When we are using a pointer to a single element of the multi-dimensional array, the pointer arithmetic
depends on the physical representation. Say, the pointer is currently pointing to the element at row-0,
column-0 of a two-dimensional array and we want it to point to the element at row-1, column-0. To
move the pointer to this new location, we need to add the total column-count of the array to the pointer.
Adding the column-count moves the pointer to same column of the next row. When we are at the last
column of a row and do an increment of the pointer, the pointer moves to the first column of the next
row.

When we are using a row-pointer (a pointer pointing to an entire row of the array) [refer section 8.4.2],
incrementing the pointer makes it point to the next row of the array. Adding a count with the pointer
advances the pointer by the number of rows equal to the count (pointer_name + count). For example, if
we add 2 with the pointer, the pointer advances by 2 rows. Instead of advancing the pointer or adding a
count to it, we may also use the index of the row to directly make the pointer point to the required row
(pointer_name[row-index]). Now as we are pointing to the starting address of the required row, we may
just add the column index to point to the required column within the row. We may then use the
indirection operator * (asterisk) to get the value stored at that location. Assuming the pointer is
pointing to the first row (Row-0), we may do any of the following to get the address and value of an
element within the two-dimensional array.

Memory Address Value

*(pointer_name + row-index) + column-index *(*(pointer_name + row-index) + column-index)
pointer_name[row-index] + column-index *(pointer_name[row-index] + column-index)

If the ‘pointer_name’ is incremented to point to the required row of the two-dimensional array, we may
use the following syntax to get the value at a specific ‘column-index’ corresponding to the row.
(*pointer_name)[column-index]

Now, let us understand the above using a diagram. Assume, we have a two-dimensional integer array of
4 rows and 5 columns ([4][5] array). We have a pointer ‘ptr1’ which points to a single element of the
array and a pointer ‘ptr2’ which points to an entire row of the array. The integer data type takes 4 bytes
of memory space and the array starts from the imaginary memory location 100. So, the two pointers
are declared as:
int *ptr1 = &array_name[0][0];
int (*ptr2)[5] = &array_name[0];

125
Quick & Indepth C With Data Structures

As per the above diagram, various pointer operations on ‘ptr1’ and ‘ptr2’ will yield the following results:
Operation Result Explanation
ptr1 + 3 112 Address of element [0][3] as 'ptr1' is pointer to a single element /
variable and not a row
ptr1 + 5 120 Address of element [1][0], due to sequential representation in actual
physical memory
ptr1 + 17 168 Address of element [3][2]
*(ptr1 + 5) 6 Value of element [1][0]
ptr1++ 104 Increments 'ptr1' to point to the next element ‘[0][1]’
ptr2 + 2 140 Address of Row-2 (3rd row of the array, as 'ptr2' points to an entire row)
*ptr2 + 2 108 Address of element [0][2] (3rd column of row to which 'ptr2' is pointing)
*(*ptr2 + 2) 3 Value of the element [0][2]
*(ptr2 + 3) 160 Address of element [3][0]. It is same as specifying '*(ptr2 + 3) + 0'
*(ptr2 + 3) + 2 168 Address of element [3][2] (3rd column of the row pointed by 'ptr2 + 3')
*(*(ptr2 + 3) + 2) 18 Value of element [3][2]
ptr2[2] + 3 152 Address of element [2][3]
*(ptr2[2] + 3) 14 Value of element [2][3]
(*ptr2)[2] 3 Value of element [0][2]. (3rd column of the row pointed by ‘ptr2’)

In our next example, we will accept an array of numbers from the user. We will then calculate the sum
and mean of all the numbers. To understand pointers better, we will accept and enumerate the array

126
 POINTERS

elements using a pointer.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
int rgiNumbers[MAX_NUMBERS] = { 0 };
int *ptr1 = NULL;
long lSum = 0;
double dblMean = 0;

/* Accept the numbers from the user. */

for (ptr1 = rgiNumbers; (ptr1 - rgiNumbers) < MAX_NUMBERS; ptr1++)
{
printf("Enter the number: ");
scanf("%d", ptr1);

lSum = lSum + *ptr1;

}

dblMean = (double)lSum / (double)MAX_NUMBERS;

printf("\nThe mean of the numbers is %.02lf.\n\n", dblMean);

return 0;
}

In our next program we will have two matrices and two pointers. Each pointer will point to an entire row
of a matrix. We will use these two pointers to subtract one matrix from another and print the result.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define NUM_ROWS 5
#define NUM_COLS 4

int main(void)
{
/* Variable Declaration & Initialization */
int rgiMatrix1[NUM_ROWS][NUM_COLS] = { { 3, 6, 9, 12 },
{ 15, 18, 21, 24 },
{ 27, 30, 33, 36 },
{ 39, 42, 45, 48 },
{ 51, 54, 57, 60 } };

int rgiMatrix2[NUM_ROWS][NUM_COLS] = { { 2, 4, 6, 8 },
{ 10, 12, 14, 16 },
{ 18, 20, 22, 24 },
{ 26, 28, 30, 32 },
{ 34, 36, 38, 40 } };
int (*pMatrix1)[NUM_COLS] = &rgiMatrix1[0];
int (*pMatrix2)[NUM_COLS] = &rgiMatrix2[0];
int iRow = 0;
int iCol = 0;
long lDifference = 0;

printf("The difference of the matrices is:\n");

127
Quick & Indepth C With Data Structures

for (; iRow < NUM_ROWS; iRow++, pMatrix1++, pMatrix2++)

{
for (iCol = 0; iCol < NUM_COLS; )
{
lDifference = (*pMatrix1)[iCol] - (*pMatrix2)[iCol];
/* Instead, we can also do the following: */
/* lDifference = *(*pMatrix1 + iCol) - *(*pMatrix2 + iCol); */
printf("%2ld", lDifference);
if (++iCol < NUM_COLS)
{
printf(", ");
}
}

printf("\n");
}

printf("\n");

return 0;
}

From the above discussions we know that pointers can take part in various kinds of arithmetic
expressions. Some more examples are:
iValue += *ptr1;
iValue = *ptr1 * *ptr2;
iValue = *ptr1 / *ptr2;
*ptr2 += *ptr1;
ptr1 += 3;
ptr1++;
ptr2 = ptr1 + 2;

Notice in the second statement above, we have used a space between the two consecutive asterisk (*).
This is required for the compiler to understand that the first * is for multiplication and the second one is
the indirection operator. If the space is not given, the compiler will confuse it with a pointer to pointer
expression (we will learn about pointer to pointer a little later). Similarly in the third statement, we have
used a space between the forward slash (/) and the asterisk (*). Absence of a space between these two
will make the compiler assume it as the start of a comment section which starts with the character pair
‘/*’.

There are few kinds of arithmetic expressions which cannot be used with pointers. Some arithmetic
expressions involving pointer variables are not allowed as they are outright meaningless to be
performed. Addition of two pointers and pointers involved in multiplication or division are some of the
disallowed operations involving pointers. Assuming two pointers ‘ptr1’ and ‘ptr2’, some examples of
invalid use of pointers in expressions are:
ptr1 * 3;
ptr1 * ptr2;
ptr1 / 2;
ptr1 / ptr2;
ptr1 + ptr2;

128
 POINTERS

In addition to the arithmetic operations, pointers can also take part in expressions involving relational
operators. These expressions are really useful when we are working with pointers to array elements.
Expressions like ‘ptr1 < ptr2’, ‘ptr2 >= ptr1’, ‘ptr1 != ptr2’, ‘ptr1 == ptr2’, ‘ptr1 != NULL’ and ‘ptr2 ==
NULL’ are valid expressions involving pointers. As discussed in section 7.7, the last two expressions
‘ptr1 != NULL’ and ‘ptr2 == NULL’ may also be written as ‘ptr1’ and ‘!ptr2’ respectively.

8.6 POINTER TO ANOTHER POINTER

Just like pointers point to the memory addresses of other variables, they can also point to the memory
locations of other pointers. We have learned in section 8.2 that pointers are also variables which occupy
memory space and have an address of their own. We can declare another pointer which points to this
memory address. In other words the second pointer will have the address of the first pointer as its
value. The syntax for declaring a pointer to another pointer is:
Data-Type **pointer_name;
Notice, we have two asterisk (*) in the declaration of a pointer to pointer. The number of asterisk in the
declaration is proportional to the level of the pointer. This means that a pointer to a simple variable will
have a single asterisk, a pointer to another pointer will have two asterisks, and a pointer to a pointer
which again points to another pointer will have three asterisks and so on. We rarely require pointers
with more than two levels which we will explore in this section.

In the above image, we can see a pointer ‘pointer_name1’ is pointing to the variable ‘iValue’ and
contains the memory address of ‘iValue’ as its value. Similarly, a pointer ‘pointer_name2’ is pointing to
the pointer variable ‘pointer_name1’ and contains the memory address of ‘pointer_name1’ as its value.

Similar to other variables, the ‘address of’ operator (&) is used to assign the address of a pointer to
another pointer.
pointer_name2 = &pointer_name1;
The indirection operator (*) is used with ‘pointer_name2’ to retrieve the value stored in ‘pointer_name1’.
Using a single indirection operator will get us the value stored in ‘pointer_name1’ which as per our
above diagram will be the value 100 (the address of iValue). To get the value stored in ‘iValue’, we have
to use two indirection operators with ‘pointer_name2’.

129
Quick & Indepth C With Data Structures

Expression Value Explanation

pointer_name1 100 Address of iValue (The value stored in pointer_name1).
*pointer_name1 978562 Value stored in iValue.
pointer_name2 656 Address of pointer_name1 (The value stored in pointer_name2).
*pointer_name2 100 Address of iValue (The value stored in pointer_name1).
**pointer_name2 978562 Value stored in iValue.

So each indirection operator gets the value stored in the location specified at the right side of the
operator. When we are specifying ‘*pointer_name2’, it retrieves the value stored in the address location
pointed by ‘pointer_name2’. The address location pointed by ‘pointer_name2’ is 656 (the address of
pointer_name1). The value stored in this location i.e. at ‘pointer_name1’ is 100 (the address of iValue1).
So, ‘*pointer_name2’ returns the value 100.
When we are specifying ‘**pointer_name2’, it retrieves the value stored in the address location pointed
by ‘*pointer_name2’. From our previous finding, ‘*pointer_name2’ is 100. This makes ‘**pointer_name2’
give us the value stored in the address location 100 (the value stored in ‘iValue’). So, ‘**pointer_name2’
returns the value 978562.

In our next program we will have an integer array of 10 elements. We will declare a pointer ‘ptr1’ which
points to the first element of the array. We will have another pointer ‘ptr2’ which points to the first
pointer ‘ptr1’. We will then calculate the sum of all the elements of the array by using the second
pointer ‘ptr2’. We will increment the first pointer ‘ptr1’ until it crosses the address of the last element of
the array. To make matters more complicated, we will do the increment and the addition using the
pointer ‘ptr2’. This is a very cryptic way of solving a very simple problem. We are just doing it to get a
better understanding of the concept of pointer to another pointer.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
int rgiValues[10] = { 7, 72, 83, 20, 37, 6, 10, 26, 9, 57 };
int *ptr1 = &rgiValues[0];
int **ptr2 = &ptr1;
int iSum = 0;

/* We are incrementing the value at the location pointed by ptr2. As ptr2 is pointing
to ptr1, we are essentially incrementing ptr1 to point to the next element. We are
not changing ptr2 in any way. We are incrementing ptr1 until it points to an address
greater than the last element of the array which is 'rgiValues[9]'. *ptr2 refers
to the value at the location pointed by ptr2, which in other words is ptr1. */
for (; *ptr2 <= &rgiValues[9]; (*ptr2)++)
{
iSum += **ptr2;
}

/* Print the sum. */

printf("The sum of the elements is %d.\n\n", iSum);

return 0;
}

130
 POINTERS

Output
The sum of the elements is 327.

8.7 SAMPLE PROGRAMS

1. In our next program we will left-rotate an array. The number of places the element values will
be rotated is user specified. So, if the array is {'A', 'B', 'C', 'D', 'E'} and we left-rotate 3 places,
the array will become {'D', 'E', 'A', 'B', 'C'}.

/* Header Files */
#include <stdio.h>

/* Defined Values */
/* ARRAY_SIZE must be greater than 1. */
#define ARRAY_SIZE 10

int main(void)
{
/* Variable declaration and initialization. */
char rgValues[ARRAY_SIZE] = { 0 };
char *ptrCur = NULL;
char chFirst = '\0';
int iNumRotation = 0;
int iCurRotation = 0;

/* Accept all the characters from the user.

To understand pointers better, instead of an index counter,
we are using a pointer to loop through the array elements. */
for (ptrCur = rgValues; ptrCur < (rgValues + ARRAY_SIZE); ptrCur++)
{
printf("Enter the element rgValues[%d]: ", (ptrCur – rgValues));
scanf("%c", ptrCur);
}

printf("\n");
while (iNumRotation <= 0 || iNumRotation >= ARRAY_SIZE)
{
printf("Enter the number of places to rotate: ");
scanf("%d", &iNumRotation);
}

/* Print the array before the rotation.

To understand pointers better, instead of an index counter,
we are using a pointer to loop through the array elements. */
printf("\n\nThe array before the rotation is:\n");
for (ptrCur = rgValues; ptrCur < (rgValues + ARRAY_SIZE); ptrCur++)
{
printf("%c", *ptrCur);
if (ptrCur < (rgValues + (ARRAY_SIZE – 1)))
{
/* We do not print the comma after the last element. */
printf(", ");
}
}

/* Rotate the array. For each rotation shift all the

array elements one place to the left. */

131
Quick & Indepth C With Data Structures

for (iCurRotation = 0; iCurRotation < iNumRotation; iCurRotation++)

{
/* Save the first array element, before it is overwritten. */
chFirst = rgValues[0];

/* Shift all the array elements one place to the left. */

for (ptrCur = rgValues; (ptrCur + 1) < (rgValues + ARRAY_SIZE); ptrCur++)
{
*ptrCur = *(ptrCur + 1);
}

/* The first array element becomes the last element after each rotation. */
*(rgValues + ARRAY_SIZE - 1) = chFirst;
}

/* Print the array after the rotation. */

printf("\n\nThe array after the rotation is:\n");
for (ptrCur = rgValues; ptrCur < (rgValues + ARRAY_SIZE); ptrCur++)
{
printf("%c", *ptrCur);
if (ptrCur < (rgValues + (ARRAY_SIZE – 1)))
{
/* We do not print the comma after the last element. */
printf(", ");
}
}

printf("\n\n");

return 0;
}

2. In our next program we will do selection sort on an array of numbers. The selection sort
algorithm sorts an array by repeatedly finding the minimum element (considering ascending
order) from unsorted part and putting it at the beginning. The algorithm maintains two
imaginary sub-arrays within a given array – The sub-array which is already sorted at the left of
the array, and remaining sub-array in the right which is unsorted.
In every iteration of selection sort, the minimum element (considering ascending order) from
the unsorted sub-array is picked and swapped with the initial element in the unsorted sub-
array. Initially, the sorted sub-array is empty and the unsorted sub-array is the entire input list.
The algorithm proceeds by finding the smallest (or largest, in case descending sort order)
element in the unsorted sub-array, swapping it with the leftmost unsorted element say 'A'. We
then continue the process from the element after this element 'A' i.e. from 'A + 1'. Assume we
have the array as 5, 7, 8, 2. In our first iteration, the entire array is unsorted. The minimum
value '2' within the array is found and swapped with '5' resulting in the array becoming 2, 7, 8,
5. Now in our next iteration, we start from '7' and try to find the next minimum value. We find '5'
which is swapped with the value '7' resulting in the array becoming 2, 5, 8, 7. Now, start from
'8' and try to find the next minimum, which is '7'. We swap '7' and '8' resulting in the array
becoming 2, 5, 7, 8, the sorted array.

/* Header Files */
#include <stdio.h>

/* Defined Values */

132
 POINTERS

#define MAX_NUMBERS 10

int main(void)
{
/* Variable Declaration & Initialization */
int rgiNumbers[MAX_NUMBERS] = { 0 };
int *ptr1 = NULL;
int *ptr2 = NULL;
int *pMinimum = NULL;
int iSwap = 0;

/* Accept the numbers from the user. */

for (ptr1 = rgiNumbers; (ptr1 - rgiNumbers) < MAX_NUMBERS; ptr1++)
{
printf("Enter the number: ");
scanf("%d", ptr1);
}

/* Now, sort the numbers. We go till second last element. We do not iterate till the last
element because when we are iterating the second last element, all elements till that
element are already sorted and have values less than or equal to the second last and the
last elements. Only the second last and last elements need sorting between themselves
which is done during the second last iteration itself i.e. if the last element is found
to have a value less than the second last element, their values are swapped. */
for (ptr1 = rgiNumbers; (ptr1 - rgiNumbers) < (MAX_NUMBERS - 1); ptr1++)
{
/* We start assuming that we have no elements with value less than this element. If
we find any, we will mark that as minimum. */
pMinimum = ptr1;

/* All elements before 'ptr1' are already sorted and have values less than
or equal to the elements which are yet to be sorted. Now, try to find the
next minimum value within the remaining unsorted elements.*/
for (ptr2 = ptr1 + 1; (ptr2 - rgiNumbers) < MAX_NUMBERS; ptr2++)
{
if (*ptr2 < *pMinimum)
{
/* We found a new minimum value. */
pMinimum = ptr2;
}
}

/* Swap the 'ptr1' element's content with the element with the least value
within the unsorted elements. If 'pMinimum' is the same as 'ptr1' i.e.
'ptr1' is the minimum, we do not need to swap. */
if (ptr1 != pMinimum)
{
iSwap = *pMinimum;
*pMinimum = *ptr1;
*ptr1 = iSwap;
}
}

/* Print the sorted sequence. We do not print the comma after the last number.*/
printf("\n\nSorted list: ");
for (ptr1 = rgiNumbers; (ptr1 - rgiNumbers) < MAX_NUMBERS; ptr1++)
{
printf("%d%s", *ptr1, ((MAX_NUMBERS - 1) == (ptr1 - rgiNumbers)) ? "\n" : ", ");
}

return 0;
}

133
Quick & Indepth C With Data Structures

3. In our next example, we will accept an array of numbers from the user. We will then calculate
the sum and mean of all the numbers. To understand pointers better, we will accept and
enumerate the array elements using a pointer.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_NUMBERS 10

int main(void)
{
/* Variable Declaration & Initialization */
int rgiNumbers[MAX_NUMBERS] = { 0 };
int *ptr1 = NULL;
long lSum = 0;
double dblMean = 0;

/* Accept the numbers from the user. */

for (ptr1 = rgiNumbers; (ptr1 - rgiNumbers) < MAX_NUMBERS; ptr1++)
{
printf("Enter the number: ");
scanf("%d", ptr1);

lSum = lSum + *ptr1;

}

dblMean = (double)lSum / (double)MAX_NUMBERS;

printf("\nThe mean of the numbers is %.02lf.\n\n", dblMean);

return 0;
}

4. Let us write a program which declares and initializes an array. It accepts a number from the
user and a location within the array. It inserts the number in the location specified by the user.
We need to move all elements on the right of the specified location by one place to make space
for the new number. The number at the last element within the array is simply dropped off.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_NUMBERS 10

int main(void)
{
/* Variable Declaration & Initialization */
int rgiNumbers[MAX_NUMBERS] = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 };
int *ptrLoc = NULL;
int *ptrMove = NULL;
int iNewNumber = 0;
int iLocation = -1;

/* Accept the number to insert. */

printf("Enter the number to insert: ");
scanf("%d", &iNewNumber);

134
 POINTERS

/* Accept the location where to insert this number within the array. */
while (iLocation < 0 || iLocation >= MAX_NUMBERS)
{
printf("Enter its location (0-%d): ", MAX_NUMBERS – 1);
scanf("%d", &iLocation);
}

/* Starting from the array end to the specified location, move all
the elements one place to the right. The last array element has
nowhere to go and so, gets overwritten by the previous element. */

/* Locate the position where we will do the insert. */

ptrLoc = rgiNumbers + iLocation;

/* Move all the elements to the right by one place. */

ptrMove = rgiNumbers + (MAX_NUMBERS – 1);
for (; ptrMove > ptrLoc; ptrMove--)
{
*ptrMove = *(ptrMove – 1);
}

*ptrLoc = iNewNumber;

/* Now, print the array. We do not print the comma after the last number. */
printf("\n\nRevised array: ");
for (ptrMove = rgiNumbers; (ptrMove - rgiNumbers) < MAX_NUMBERS; ptrMove++)
{
printf("%d%s", *ptrMove,
((MAX_NUMBERS - 1) == (ptrMove - rgiNumbers)) ? "\n\n" : ", ");
}

return 0;
}

135
Quick & Indepth C With Data Structures

CHARACTER STRING OPERATIONS

[CHAPTER-9]

9.1 INTRODUCTION
We have read about strings in our previous chapters. To reiterate, string is an array of characters
defined between double quotation marks. A ‘\0’ (NULL) character (ASCII value 0) marks the end of the
string. A character array or constant which does not end with the NULL character is not a string but is
simply an array of characters. Consider the following examples:
1. char szString1[] = “String 1”;
2. char szString2[20] = “String 2”;
3. const char *pszString3 = “String 3”;
4. char szString4[] = { ‘S’, ‘t’, ‘r’, ‘i’, ‘n’, ‘g’, ‘ ‘, ‘4’, ‘\0’ };
5. char szString5[20] = { ‘S’, ‘t’, ‘r’, ‘i’, ‘n’, ‘g’, ‘ ‘, ‘5’, ‘\0’ };
6. char szString6[20] = { 65, 66, 67, 68, 69, 0 };
7. char rgArray1[] = { ‘N’, ‘o’, ‘t’, ‘ ’, ‘a’, ‘ ’, ‘s’, ‘t’, ‘r’, ‘i’, ‘n’, ‘g’, ‘1’ };
8. char rgArray2[20] = { ‘N’, ‘o’, ‘t’, ‘ ’, ‘a’, ‘ ’, ‘s’, ‘t’, ‘r’, ‘i’, ‘n’, ‘g’, ‘2’ };
9. char rgArray3[20] = { 65, 66, 67, 68, 69 };
10. static char szString7[20] = { 65, 66, 67, 68, 69 };

Let us have a look at what happens in the above examples.

The first example declares a character array of 9 elements and initializes it to the string ‘String 1’. The
first element contains the character ‘S’, the second element contains the character ‘t’, the third
contains the character ‘r’ and so on. The last element contains the character ‘\0’ (NULL). We have
enclosed the term “String 1” within double quotes indicating to the compiler that it needs to be treated
as a string, and so should end with the ‘\0’ (NULL) character. During compilation, the compiler
automatically calculates the number of elements needed to store the data (specified on the right)
including the terminating ‘\0’ (NULL) character. The compiler sets the size of the character array
‘szString1’ as per this calculated size. As the string on the right “String 1” has 8 characters, the
compiler sets the size of the character array ‘szString1’ to 9, where the first 8 elements will contain the
8 specified characters and the last element will have the terminating ‘\0’ (NULL) character.

The second example is very similar to the first, except for the fact that the character array ‘szString2’
contains 20 elements instead of the 9 for ‘szString1’. This is because we have explicitly stated the size
of the array ‘szString2’ to be of 20 elements. So in this case, the compiler does not automatically assign
the array size by calculating the size requirement.

The third example declares a string “String 3” in the memory. As a string is nothing but a character
array ending with the ‘\0’ (NULL) character, the program allocates a character array somewhere in
memory and assigns the string “String 3” to it. This array ends with the ‘\0’ (NULL) character. The

136
 CHARACTER STRING OPERATIONS

program declares a character pointer ‘pszString3’ and assigns the starting address of the array to it. We
can then use this pointer to access this string within our program. As we have declared the pointer
‘pszString3’ as a constant (using the keyword const), the pointer cannot be changed anywhere within
the program. It can only be read or accessed, but cannot be modified. As the pointer is the only way for
us to access the string, modifying this pointer will make the string constant inaccessible. So, we have
declared the pointer as a const.

In our fourth example, we are not initializing the array using a string literal (specified within double
quotes) as in the previous cases. We are specifying the individual characters to be stored in each of the
array elements. So, the compiler does not automatically append a ‘\0’ (NULL) character at the end of
the array. So, we are explicitly specifying a ‘\0’ character as the last value during the initialization. The
compiler automatically calculates the required size of the array equal to the number of specified values
during the initialization which is 9 (including the ending ‘\0’ character). As we have specified a ‘\0’ at
the end, the array ‘szString4’ has a string “String 4” as its content.

The fifth example is very similar to the fourth, except for the fact that the character array ‘szString5’
contains 20 elements instead of the 9 for ‘szString4’. This is because we have explicitly stated the size
of the array ‘szString5’ to be of 20 elements.

The sixth example is again very similar to the fifth. The only difference is that we have initialized the
array elements by directly using ASCII values instead of characters. The array ‘szString6’ will have the
string “ABCDE” as the ASCII value 65 refers to the character ‘A’, the value 66 refers to the character ‘B’
and so on. We have specified 0 at the end which is the ASCII value of the ‘\0’ (NULL) character.

In our seventh example the array ‘rgArray1’ does not have a string as its content. We have neither
initialized it using a string literal (specified within double quotes) nor we have specified a ‘\0’ character
at the end during initialization. The array ‘rgArray1’ will have 13 elements i.e. the number of values
specified during the initialization. It will have no element left for the terminating ‘\0’ character.

The next example is similar to the seventh except for the fact that the array contains 20 elements
instead of 13. This is because we have explicitly specified 20 as the array size.

The ninth example too is very similar to the eighth. The only difference is that we have initialized the
array elements by directly using ASCII values of the characters. This array too does not contain a string
as it does not end with a ‘\0’ character.

Our last example is an interesting one. We have used the storage qualified static for the array
‘szString7’. From our previous understanding in section 2.9, we know that uninitialized static variables
are automatically initialized to the value 0. So in this case too, all our uninitialized array elements will
have the value 0. We have 20 elements in our array and we are initializing the first 5 elements ASCII
values 65, 66, 67, 68 and 69 corresponding to the characters ‘A’, ‘B’, ‘C’, ‘D’, ‘E’ respectively. As the
array is declared as static, the rest 15 uninitialized elements will automatically be initialized to the value
0. So, this makes our array contain the string “ABCDE” as its content.

137
Quick & Indepth C With Data Structures

The common operations performed on character strings are:

• Reading and writing strings
• Copying one string to another
• Combining strings
• Comparing two strings for equality
• Extracting a sub-string (portion) from a larger string
• Converting a string with numeric characters to an int, float or double data type

9.2 UNICODE CHARACTERS

As we have read in section 2.2, a new character representation called Unicode is defined which is
capable of representing characters from most of the World languages, unlike ASCII which is capable of
representing only English characters. Till now, we have read about ASCII character representation in
our earlier chapters. In this chapter we will study a bit about working with Unicode characters.

The first 128 characters of ASCII are also present in Unicode. So, the first 128 symbols are the same in
ASCII and Unicode. Unicode can be implemented by different character encoding mechanisms.
Character encoding means the representation of a character in digital format. The Unicode standard
defines UTF-8, UCS-2, UTF-16, UTF-32 and many other character encodings.

UTF-8 encoding uses one byte for the first 128 characters, and up to 4 bytes for other characters. So, it
is a variable size encoding scheme which varies depending on the character used. The first 128
characters of UTF-8 are the same as ASCII and are represented using 1 byte which makes ASCII and
UTF-8 same for these 128 characters.

UCS-2, a precursor of UTF-16 simply uses two bytes (16 bits) for each character but can only encode
65,536 characters. UCS-2 is widely used across software and is also the one supported by many
operating systems and C compilers. As it represents only 65,536 characters, not all World language
characters can be accommodated in it. UTF-16 extends UCS-2, by using the same 16-bit encoding as
UCS-2 for most languages, and a 4-byte encoding for others.

UTF-32 (also referred to as UCS-4) uses four bytes for each character. Like UCS-2, the number of bytes
per character is fixed. UTF-32 is capable of representing all World characters. However, because each
character uses four bytes, UTF-32 uses significantly more memory than other encoding mechanisms
and so, and is not widely used.

ASCII characters are represented internally using 7 bits. ASCII was very limited in capability and could
support only 128 characters. This was sufficient for the English language but, could not represent most
other world languages. Another standard ANSI was developed which used all of the 8 bits of a byte. For
compatibility purposes, the first 128 characters of ANSI was the same as ASCII. Due to the use of 1
extra bit, ANSI supported different code pages which allowed it to represent various character sets as
defined by different vendors. Each vendor could define a specific code page to support a certain
language and use that code page within its programs. A Code Page (also referred to as Character Set or

138
 CHARACTER STRING OPERATIONS

Encoding) is a table of values where each character has been assigned a numerical representation and
that numerical value is used to identify that character. ANSI too faced a big limitation. Due to the use of
different code pages across different vendors, there was no standard which defined all of them. So, a
program written for one computer may not work properly when executed on another computer. Also,
ANSI still could not support most world languages.

Unicode was defined to overcome all the limitations of ASCII and ANSI. Due to the use of multiple bytes
to represent one character, it extended the reach of Unicode by many folds. Unicode is no longer limited
like ASCII and ANSI and could represent almost all world languages. Unlike ANSI, Unicode is controlled
by a single standard which allows the programs written for Unicode to be easily ported from one
computer to another.

As Unicode uses multiple bytes to represent one character, characters encoded using any of the
Unicode encodings are called wide characters. This is because of the availability of a wide range of
values to represent the characters.

Another type of characters called multi-byte characters are defined by the use of one or more bytes for
a single character (typically 1 or 2 bytes). Single byte characters are a special cases for multi-byte
character set and many multi-byte character set define ASCII as a subset of them.

As ASCII characters are represented in C using the data type 'char', a new type 'wchar_t' is defined
which represents a Unicode character. Unlike the data type 'char' whose size is 1 byte across compilers,
the size of 'wchar_t' varies between compilers depending on the encoding mechanism used internally by
the compilers. 'wchar_t' has a minimum size of 1 byte and can go up to 4 bytes, with 2 bytes as the
most widely used across compilers.

We have seen that ASCII character constants are represented within single quotes ('A', '1', 'z', '#', ...)
and ASCII string constants are represented within double quotes (“C Language”, “Tania”, …). In case of
Unicode, we prefix a L before the character or string constant. So, the character 'A' will be written as
L'A' and the string “C Language” will be written as L”C Language” in Unicode. Some example usage are:
wchar_t wchValue = L'w';
wchar_t wchValue = 65; /* The ASCII value of 'A' */
wchar_t wchValue = L'5';
wchar_t wszName[] = L”Tania”;
wchar_t wszName[20] = L”C Language”;
wchar_t wszName[3] = { L'W', L'e', L'\0' }; /* The string “We” */
wchar_t *pwszName = L“Williamson”;

9.3 ACCEPTING STRINGS FROM USER

We have read about input and output operations on strings in chapters 4 and 5. From these chapters
we know the behavior of the functions 'getchar', 'scanf' and 'gets'.

'scanf' function is used to receive formatted input from the user. Character strings are accepted using

139
Quick & Indepth C With Data Structures

the '%s' format specifier. Example:

char szName[20] = { 0 };
scanf(“%s”, szName);

When we are accepting strings using the 'scanf' function, the & (ampersand) sign is not required before
the variable name. We will read more about this in our chapter on pointers. 'scanf' automatically
terminates the string that is read using a '\0' (NULL) character. So when declaring the size of the array,
we should be careful in accounting for this NULL character at the end.

The problem with 'scanf' function is that it does not check overflow conditions i.e. if the user has
entered a string larger than what our array can hold. To address this, we need to use the field width
(please read section 4.3) when accepting the input, as given below:
scanf(“%19s”, szName);

The above statement will only read 19 characters into the string 'szName', though it will keep accepting
from the user until it encounters any white space (space, tab, carriage return, new line). So, even if we
enter a string of 30 characters long, the function will copy only the first 19 characters into the string
and terminate it with a '\0' (NULL) character, dropping rest of the 11 characters.

More secured version of 'scanf' has been created and are supported by some of the latest compilers.
The function 'scanf_s' is similar to 'scanf' except that we need to specify the array/buffer length too,
when we are accepting a string input. The function has the following declaration:
int scanf_s( const char * format, ... );

This protects the program from overflow conditions even when no field width is specified. The buffer
length is required only when a string or character value is accepted as input. For other variables, no
length needs to be specified. The buffer length should include the space for the terminating NULL
character. We may also specify the field width. If no field width is specified, and the string read is too big
to fit in the array buffer, nothing is written to that buffer. Consider the following example:
char szName[20] = { 0 };
scanf_s(“%s”, szName, 20);

Providing a hard-coded value (20 in our above example) is never recommended and must be avoided. If
in the future, we change the array length to something else and forget to do the same within the
'scanf_s' function, the program may generate errors or even crash. To avoid this, we may do three
things:
• Define a symbolic name using the '#define' statement and use that name (in place of the hard-
coded value) in both the places. Changing the value within the '#define' statement will have the
effect in both the places at one go. Consider the following example:
#define MAX_SIZE 20
char szName[MAX_SIZE] = { 0 };
scanf_s(“%s”, szName, MAX_SIZE);

140
 CHARACTER STRING OPERATIONS

• Rather than specifying the hard-coded value of 20 in the 'scanf_s' function, we can specify the
array length by making the compiler compute it during compile time. It is done using the macro
'_countof'.
char szName[20] = { 0 };
scanf_s(“%s”, szName, _countof(szName));

• If the compiler with which we are working does not support the '_countof' macro, we can define
a macro which does the same as what '_countof' does. Refer to section 2.10 for information on
defining macros.
#define ARRAYSIZE(x) sizeof(x) / sizeof(x[0])
char szName[20] = { 0 };
scanf_s(“%s”, szName, ARRAYSIZE(szName));

The macro call 'ARRAYSIZE(szName)' is replaced by the expression corresponding to the macro
and it becomes 'sizeof(szName) / sizeof(szName[0])'. The macro calculates the number of
elements in the array 'szName'. It divides the size (memory usage) of the entire array
'sizeof(szName)' by the size (memory usage) of one single element of the array
'sizeof(szName[0])'. So, in effect we get the total number of elements within the array.

Another problem with 'scanf' function when working with character strings is that it terminates its input
on the first white space (space, tab, carriage returns, new line) it encounters. So, when we type “David
Miller” as its input, 'scanf' will only accept “David” as the input, because of the presence of a space
between the words "David" and "Miller". To accept both the words, we may use two character arrays
within the 'scanf' function.
char szFirstname[11] = { 0 };
char szLastname[11] = { 0 };
scanf(“%s %s”, szFirstName, szLastName);

The above example will work only when we are sure that there will be two separate words to be entered.
For all other scenarios, the above example will fail. So, wherever a string (to be accepted) may contain
white spaces in between, we should avoid using the 'scanf' function and rely on another input function
‘fgets’ with ‘stdin’ as the input stream.

All the above input methods display the text entered by the user. This is not suitable for situations
where we are looking to accept secret information like password. We can use the 'getch' function for
accepting such information. We will accept one character at a time within a loop. A sample program is
given below.

Program for accepting a password from the user:

/* Header Files */
#include <stdio.h>
#include <conio.h>

/* Defined Values */

141
Quick & Indepth C With Data Structures

#define MAX_PASS 20

int main(void)
{
/* Variable declaration and initialization. */
char szPassword[MAX_PASS + 1] = { 0 }; /* One extra for NULL. */
int idx = 0;

/* Accept the password in a loop. */

printf("Enter the password: ");
while (idx < MAX_PASS)
{
szPassword[idx] = getch();
if (szPassword[idx] == '\n' || szPassword[idx] == '\r')
{
break;
}

idx++;
}

/* The terminating ‘\0’ character to make it a string. */

szPassword[idx] = '\0';

return 0;
}

The program accepts the entire password within a loop. It keeps accepting till the number of characters
entered is less than the array length. If the <Return> (Enter) key is hit, it breaks out from the 'while'
loop.

9.4 PRINTING STRINGS TO CONSOLE

We have read about formatted output using the 'printf' function in the section 4.4.2. Here we will go
through some more examples of formatted output of strings. We know character strings are printed
using the '%s' format specifier. We should remember that the strings must be '\0' (NULL) terminated.
We may use flags, field width and precision along with the conversion character as specified in section
4.3. In the following example, the field width is 6 and the precision is 4.
printf(“%6.4s”, “Jupiter”);

Some of the rules which apply when printing a string using the format specifiers are:
• The number of characters printed is equal to the precision specified or the length of the string
whichever is lower.
• We can left-align the printed text by specifying a '-' (minus) sign to the immediate right of the
'%' sign.
• If the printed text is right-aligned (the default), spaces equal to the difference between the field
width and the precision is printed on the left side of the printed text. So, in the above example 2
spaces will be printed to the left of the word “Jupi”.
• If the printed text is left-aligned, spaces equal to the difference between the field width and the
precision is printed on the right side of the printed text. So, in the above example 2 spaces will
be printed to the right of the word “Jupi”, if it was made left-aligned.
• If no precision is specified and the specified field width is less than the length of the string then

142
 CHARACTER STRING OPERATIONS

the entire string is printed.

• If the precision is specified as 0 (zero), nothing is printed.

Let us write a program which prints the words “C Language” using various format specifiers.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable declaration and initialization. */
char szName[] = "C Language";

printf("************\n");
printf("%s\n", szName);
printf("%12s\n", szName);
printf("%-12s\n", szName);
printf("%12.6s\n", szName);
printf("%-12.6s\n", szName);
printf("%.0s\n", szName);
printf("%.6s\n", szName);
printf("%12.0s\n", szName);
printf("************\n\n");

return 0;
}

Output:

************
C Language
C Language
C Language
C Lang
C Lang

C Lang

************

9.5 ARITHMETIC OPERATIONS WITH CHARACTERS

We can perform arithmetic operations with characters in the same way we do with integers. Whenever a
character appears in an arithmetic expression, it is converted to an integer value depending on the
character representation we are using. If we are using ASCII, the character is converted to its
equivalent ASCII value before performing the arithmetic operation. If we are using Unicode, the
character is converted to its equivalent Unicode value. Consider the following expressions:

Expression Explanation
int iValue = ‘A’; The ASCII value of ‘A’ i.e. 65 is assigned to the variable ‘iValue’.
int iValue = ‘Z’ - ‘A’ + 1; The ASCII value of ‘A’ is subtracted from the ASCII value of ‘Z’
resulting in the value 25 (90 – 65). 1 is then added with 25 to get

143
Quick & Indepth C With Data Structures

Expression Explanation
26, which is assigned to ‘iValue’.
char ch = ‘B’; Prints the string “It is UPPERCASE.”.
if (ch >= ‘A’ and ch <= ‘Z’)
printf(“It is UPPERCASE.”);
else
printf(“It is lowercase.”);

9.6 ARITHMETIC OPERATIONS WITH STRINGS

Unlike single characters which can take part in arithmetic operations, strings being arrays cannot take
part in arithmetic operations. Neither two strings can be concatenated together by simply adding them
together nor they can be compared for equality. Consider some of the following invalid usages:

char szString1[] = “I Love ”;

char szString2[] = “C Language”;
char szString3[50] = szString1 + szString2; /* INVALID */
char szString1[] = “A String”;
char szString2[] = “A String”;
if (szString1 == szString2) /* INVALID */
printf(“The strings are equal”);
char szString1[] = “A String”;
char szString2[] = “A String”;
if (szString1 > szString2) /* INVALID */
printf(“szString1 is greater than szString2”);

In the above examples, the statements in bold are invalid. We cannot concatenate two strings together
by simply adding them. We need to use specific functions for the purpose which we will discuss later in
this chapter. In the second example, we are comparing two strings for equality using the relational
operator ‘==’. We cannot compare two strings using any of the relational operators like ‘==’, ‘!=’, ‘>’,
‘<’. We have specific functions for the purpose which we are going to read a little later in this chapter.

9.7 STRING PROCESSING FUNCTIONS

Most standard string processing functions are declared in the header file ‘string.h’ for the ANSI version
and ‘wchar.h’ for the wide character version. So when using any of the following functions, we must
include the respective header file within our program.

9.7.1 GENERAL STRING PROCESSING FUNCTIONS

• ANSI → size_t strlen(const char *str);

Wide Character → size_t wcslen(const wchar_t *wstr);

144
 CHARACTER STRING OPERATIONS

◦ Returns the length of the string ‘str’ provided as an argument to the function. It accepts the
pointer to the string and returns the number of characters in the string excluding the
terminating ‘\0’ (NULL) character. The return type is ‘size_t’ which is a type-defined type
and is compiler dependent. It must always be of unsigned integer type. Usually it is
‘unsigned int’ in 32-bit programs and ‘unsigned _int64’ in 64-bit programs. The length of
the string is calculated as the number of characters from the start of the string to the
terminating ‘\0’ (NULL) character (excluding the terminating ‘\0’ (NULL) character itself in
the count). Usage example:
char szLanguage[] = “C Language”;
size_t cchLen = strlen(szLanguage); /* Returns 10 */
wchar_t szLanguage[] = L“C Language”;
size_t cchLen = wcslen(szLanguage); /* Returns 10 */

• ANSI → char * strcpy(char *destination, const char *source);

Wide Character → wchar_t * wcscpy(wchar_t *destination, const wchar_t *source);
◦ Copies the string pointed by ‘source’ into the array pointed by ‘destination’, including the
terminating ‘\0’ (NULL) character. To avoid overflows, we must ensure that the 'destination'
array is able to contain all the characters in the 'source' string, including the terminating
‘\0’ (NULL) character. If it is not large enough, strcpy/wcscpy will write past the
'destination' array and will cause a program crash. The function returns the starting
address of the 'destination' array. Usage example:
char szString1[] = “C Language”;
char szString2[11];
strcpy(szString2, szString1);
wchar_t szString1[] = L“C Language”;
wchar_t szString2[11];
wcscpy(szString2, szString1);

• ANSI → char * strcat(char *destination, const char *source);

Wide Character → wchar_t * wcscat(wchar_t *destination, const wchar_t *source);
◦ Appends a copy of the 'source' string to the 'destination' string. The terminating ‘\0’ (NULL)
character in 'destination' is overwritten by the first character of 'source', and a ‘\0’ (NULL)
character is included at the end of the new string formed by the concatenation of both the
strings in 'destination'. To avoid overflows, we must ensure that the 'destination' array is
able to contain all the characters after the concatenation, including the terminating ‘\0’
(NULL) character. If it is not large enough, strcat/wcscat will write past the 'destination'
array and will cause a program crash. The function returns the starting address of the
'destination' array. Usage example:
char szString1[] = “C Language”;
char szString2[20] = “I Love ”;
strcat(szString2, szString1);
wchar_t szString1[] = L“C Language”;

145
Quick & Indepth C With Data Structures

wchar_t szString2[20] = L“I Love ”;

wcscat(szString2, szString1);

• ANSI → int strcmp(const char *str1, const char *str2);

Wide Character → int wcscmp(const wchar_t *str1, const wchar_t *str2);
◦ Compares the string 'str1' to the string 'str2'. The function starts comparing the two strings
from the first character of each string. It continues comparing a character in 'str1' with the
same indexed character in 'str2', until they differ or a terminating ‘\0’ (NULL) character is
reached in any of the two strings. It basically does a binary value comparison between the
characters of the two strings and returns a value depending on the difference of the binary
values for the first occurrence of non-identical characters in 'str1' and 'str2'. It returns a
value indicating the difference between the strings:
Less than 0 The first character that does not match has a lower value in 'str1' than
in 'str2'.
0 (Zero) Contents of both the strings are equal.
Greater than 0 The first character that does not match has a greater value in 'str1'
than in 'str2'.

Usage example:
char szString1[15] = “I am a string”;
char szString2[18] = “I Am A String”;
int iCompare = strcmp(szString1, szString2);
printf(“The result of the comparison is %d”, iCompare);
wchar_t szString1[15] = L“I am a string”;
wchar_t szString2[18] = L“I Am A String”;
int iCompare = wcscmp(szString1, szString2);
wprintf(L“The result of the comparison is %d”, iCompare);

• ANSI → int stricmp(const char *str1, const char *str2);

Wide Character → int wcsicmp(const wchar_t *str1, const wchar_t *str2);
◦ This function is not standard C function. It is provided by many modern C compilers and
may be quite useful. This function is very similar to ‘strcmp/wcscmp’ except for the fact that
it performs a case-insensitive comparison between the two input strings. So, the uppercase
character ‘A’ is treated as same as the lowercase character ‘a’. This function will return 0
(indicating equality), if it is provided with the two strings “I AM A STRING” and “I am a
string” as its arguments.

• ANSI → const char * strchr(const char *str, int iChar);

Wide Character → const wchar_t * wcschr(const wchar_t *str, int iChar);
◦ The function searches the string ‘str’ for the first occurrence of the character whose binary
value is equal to ‘iChar’. It returns a pointer to the location in ‘str’, where it found the
character ‘iChar’. If the character is not found, the function returns NULL. Usage example:

146
 CHARACTER STRING OPERATIONS

/* Find all occurrences of the character ‘a’ within the string. */

char szString[] = “Abra Cadabra”;
char *pchr = NULL;

pchr = strchr(szString, 'a');

while (pchr != NULL)
{
printf ("Found at location %d\n", pchr - szString);
pchr = strchr(pchr + 1, 'a'); /* Find the next occurrence within the string */
}
/* Find all occurrences of the wide character ‘a’ within the string. */
wchar_t szString[] = L“Abra Cadabra”;
wchar_t *pchr = NULL;

pchr = wcschr(szString, L'a');

while (pchr != NULL)
{
wprintf (L"Found at location %d\n", pchr - szString);
pchr = wcschr(pchr + 1, L'a'); /* Find the next occurrence within the string */
}

• ANSI → const char * strrchr(const char *str, int iChar);

Wide Character → const wchar_t * wcsrchr(const wchar_t *str, int iChar);
◦ This function is very similar to ‘strchr/wcschr’. The only difference is that it performs a
reverse search of the character ‘iChar’ in the string ‘str’ and hence the extra ‘r’ within the
function name. It starts searching for the character from the end of the string pointed by
‘str’. So, in effect it finds the last occurrence of the character ‘iChar’ within the string ‘str’
and returns a pointer to its location in ‘str’. The function returns NULL, if the character is
not found within the string.

• ANSI → const char * strstr(const char *str1, const char *str2);

Wide Character → const wchar_t * wcsstr(const wchar_t *str1, const wchar_t *str2);
◦ The function searches the string 'str1' for any occurence of a string similar to 'str2' and
returns a pointer to the first occurrence of 'str2' in 'str1'. If 'str1' is the string "I am a long
long string" and 'str2' is the string "long", the function will return a pointer to the first
"long" within 'str1'. So, it will return the address of the location 'str1[7]'. If ‘str2’ is not
found in ‘str1’, the function returns NULL. Usage example:
char str1[] = “I am a long long string”;
char *pstr = strstr(str1, “long”);
wchar_t str1[] = L“I am a long long string”;
wchar_t *pstr = wcsstr(str1, L“long”);

147
Quick & Indepth C With Data Structures

• ANSI → char * strtok(char *str, const char *szDelims);

Wide Character → wchar_t * wcstok(wchar_t *str, const wchar_t *szDelims);
◦ A sequence of calls to this function returns pointers to tokens within 'str', which are
sequences of contiguous characters separated by any of the characters specified within
'szDelims'. A token is a sequence of contiguous characters appearing between two
delimiters. During the first call to this function, we need to provide a string as an argument
for 'str', whose first character is used as the starting location to scan for tokens. In
subsequent calls, we specify NULL as the argument for 'str'. The function uses the position
right after the end of the last found token as the new starting location for the search for the
next token. If 'str' is the string "C: A programming language", and 'szDelims' is specified as
" :", the function will return tokens as "C", "A", "programming" and "language". Usage
example:
char str[] = “C: A programming language”;
char *pstr = NULL;

pstr = strtok(str, " :");

while(pstr != NULL)
{
printf("%s\n", pstr);
pstr = strtok(NULL, " :");
}
wchar_t str[] = L“C: A programming language”;
wchar_t *pstr = NULL;

pstr = wcstok(str, L" :");

while(pstr != NULL)
{
wprintf(L"%s\n", pstr);
pstr = wcstok(NULL, L" :");
}

• ANSI → const char * strpbrk(const char *str1, const char *str2);

Wide Character → const wchar_t * wcspbrk(const wchar_t *str1, const wchar_t *str2);
◦ The function searches 'str1' and returns a pointer to the first occurrence in 'str1' of any of
the characters that are part of 'str2'. If no matches are found, the function returns NULL.
Usage example:
char str[] = “I Love C programming language”;
char *pChr = NULL;

pChr = strpbrk(str, "aeiouAEIOU");

while (pChr != NULL)
{
printf("%c ", *pChr);

148
 CHARACTER STRING OPERATIONS

pChr = strpbrk(pChr + 1, "aeiouAEIOU");

}
wchar_t str[] = L“I Love C programming language”;
wchar_t *pChr = NULL;

pChr = wcspbrk(str, L"aeiouAEIOU");

while (pChr != NULL)
{
wprintf(L"%c ", *pChr);
pChr = wcspbrk(pChr + 1, L"aeiouAEIOU");
}

• ANSI → size_t strcspn(const char *str1, const char *str2);

Wide Character → size_t wcscspn(const wchar_t *str1, const wchar_t *str2);
◦ The function returns the index within 'str1' of the first occurrence of any of the characters
that are part of 'str2'. The function will return the length of 'str1' if none of the characters of
'str2' are found in 'str1'. Usage example:
char str[] = “It takes 2 to tango”;
size_t idx = strcspn(str, “0123456789”);
wchar_t str[] = L“It takes 2 to tango”;
size_t idx = wcscspn(str, L“0123456789”);

• ANSI → size_t strspn(const char *str1, const char *str2);

Wide Character → size_t wcsspn(const wchar_t *str1, const wchar_t *str2);
◦ The function returns the index of the first character in the string 'str1', that does not belong
to any of the characters in 'str2'. Usage example:
char str[] = “2 C or not to see”;
size_t idx = strspn(str, “0123456789”);
wchar_t str[] = L“2 C or not to see”;
size_t idx = wcsspn(str, L“0123456789”);

• ANSI → char * strerror(int errnum);

Wide Character → wchar_t * _wcserror(int errnum);
◦ Interprets the value of a C function error code given in 'errnum' and returns a string with a
message that describes the error condition.

• ANSI → char * strncpy(char *dest, const char *src, size_t cchCopy);

Wide Character → wchar_t * wcsncpy(wchar_t *dest, const wchar_t *src, size_t cchCopy);
◦ This function is very similar to 'strcpy/wcscpy' except for the fact that it copies a maximum
of 'cchCopy' characters from 'src' to 'dest'. It stops the copy process when it reaches the
'\0' character in 'src' or it has already copied 'cchCopy' number of characters, whichever
condition comes first.

149
Quick & Indepth C With Data Structures

• ANSI → char * strncat(char *dest, const char *src, size_t cchAppend);

Wide Character → wchar_t * wcsncat(wchar_t *dest, const wchar_t *src, size_t cchAppend);
◦ This function is very similar to 'strcat/wcscat' except for the fact that it appends a
maximum of 'cchAppend' characters from 'src' to 'dest'. It stops the process when it
reaches the '\0' character in 'src' or it has already appended 'cchAppend' number of
characters, whichever condition comes first.

• ANSI → int strncmp(const char *str1, const char *str2, size_t cchCompare);
Wide Character → int wcsncmp(const wchar_t *str1, const wchar_t *str2, size_t cchCompare);
◦ This function is very similar to 'strcmp/wcscmp' except for the fact that it compares a
maximum of 'cchCompare' characters. It stops comparing if it finds a '\0' character in
either 'str1' or 'str2', or it has already compared 'cchCompare' number of characters.

• ANSI → int strnicmp(const char *str1, const char *str2, size_t cchCompare);
Wide Character → int wcsnicmp(const wchar_t *str1, const wchar_t *str2, size_t cchCompare);
◦ Just like 'stricmp/wcsicmp', this function is not standard C function. This function is
provided by many modern C compilers and may be quite useful. This function is very similar
to 'stricmp/wcsicmp' except for the fact that it compares a maximum of 'cchCompare'
characters. It stops comparing if it finds a '\0' character in either 'str1' or 'str2', or it has
already compared 'cchCompare' number of characters.

9.7.2 MEMORY COPY AND COMPARISON FUNCTIONS

Apart from the string processing functions, 'string.h' has declarations for some binary data processing
functions too. They are:
• void * memcpy(void *dest, const void *src, size_t cBytes);
◦ Copies a total of 'cBytes' bytes from the location pointed to by 'src' directly to the memory
block pointed to by 'dest'. The data type of 'dest' and 'src' may differ and is irrelevant for
the function. The result is a binary copy of the data from 'src' to 'dest'. The 'src' and 'dest'
buffers must not overlap. The function does not check for any terminating '\0' (NULL)
character in 'src'. As the function does not check the availability of 'cBytes' data in 'src' or
an adequate available buffer size in 'dest', we must ensure that 'src' has atleast 'cBytes'
bytes of data to be read and 'dest' has atleast 'cBytes' bytes of available buffer. If not so,
this function will read past the actual data in 'src' or write past the available buffer in 'dest',
causing a program crash due to overflow. Usage example:
char szElemSym[15] = { 'H', '\0', '\0', 'H', 'e', '\0', 'L', 'i', '\0', 'B', 'e', '\0', 'B', '\0', '\0' };
char rgElementSymbols[5][3];

memcpy(rgElementSymbols, szElemSym, 15);

• void * memmove(void *dest, const void *src, size_t cBytes);

◦ 'memmove' is very similar to 'memcpy' except for the fact that in case of 'memmove' the
'src' and 'dest' buffers can overlap. This added flexibility makes 'memmove' work a little

150
 CHARACTER STRING OPERATIONS

slower than 'memcpy'.

• int memcmp(const void *ptr1, const void *ptr2, size_t cBytes);

◦ The function compares the first 'cBytes' bytes of the block of memory pointed by 'ptr1' to
the first 'cBytes' bytes pointed by 'ptr2'. It performs binary comparison of the data. The
function returns 0 (Zero) if they all match, greater than 0 (Zero) if the first non-identical
bytes have a greater value in 'ptr1', and less than 0 (Zero) if the first non-identical bytes
have a greater value in 'ptr2'. Unlike 'strcmp', this function does not stop comparing after
finding a '\0' character. Usage example:
char szElemSym[15] = { 'H', '\0', '\0', 'H', 'e', '\0', 'L', 'i', '\0', 'B', 'e', '\0', 'B', '\0', '\0' };
char rgElementSymbols[5][3];
int iCompare = 0;

memcpy(rgElementSymbols, szElemSym, 15);

szElemSym[4] = 'E';
iCompare = memcmp(szElemSym, rgElementSymbols, 15);

• const void * memchr(const void *ptr, int iValue, size_t cBytes);

◦ The function searches within the first 'cBytes' bytes of the block of memory pointed by 'ptr'
for the first occurrence of 'iValue', and returns a pointer to the location in 'ptr' where it
found the value. If the function does not find the value, it returns NULL. Usage example:
char *pch = NULL;
char szString[] = "I Love to see C";
pch = (char *) memchr(szString, 's', strlen(szString));
if (pch)
printf ("Found at index %d.\n", pch – szString);
else
printf ("Not found.\n");

• void * memset(void *ptr, int iValue, size_t cBytes);

◦ The function sets the first 'cBytes' bytes of the block of memory pointed by 'ptr' to the value
'iValue'. Usage example:
char szString[] = "I Love to see C";
memset(szString, '*', 9);

9.7.3 SINGLE CHARACTER PROCESSING FUNCTIONS

Apart from the above functions declared in ‘string.h’, there are some single-character handling
functions which may be very handy when writing programs. These functions are declared in the header
file ‘ctype.h’ for the ANSI version and ‘wctype.h’ for the wide character version. All of these functions
accept an integer argument which is the binary equivalent of the character to process. The wide
character version of the functions use a user-defined type called ‘wint_t’. This type varies between
compilers and is generally of the type ‘int’. They also return an integer as their result.

151
Quick & Indepth C With Data Structures

Function Description
int tolower(int ch); Returns the lowercase equivalent of the given character ‘ch’. If no such
wint_t towlower(wint_t ch); conversion is possible, the function returns ‘ch’ itself. The function
returns the value as ‘int/wint_t’ which can be casted to ‘char/wchar_t’.
int toupper(int ch); Returns the UPPERCASE equivalent of the given character ‘ch’. If no such
wint_t towupper(wint_t ch); conversion is possible, the function returns the ‘ch’ itself. The function
returns the value as ‘int/wint_t’ which can be casted to ‘char/wchar_t’.
int isalpha(int ch); Checks whether 'ch' is an alphabet. It returns a non-zero value if ‘ch’ is
int iswalpha(wint_t ch); an alphabetic character, else returns 0 (Zero).
int isdigit(int ch); Checks whether 'ch' is a decimal digit character. It returns a non-zero
int iswdigit(wint_t ch); value if ‘ch’ is indeed a decimal digit character, else returns 0 (Zero).
int isalnum(int ch); Checks whether 'ch' is either a decimal digit or an alphabet. It returns a
int iswalnum(wint_t ch); non-zero value if ‘ch’ is an alphanumeric character, else returns 0 (Zero).
int isupper(int ch); Checks whether 'ch' is an UPPERCASE alphabet. It returns a non-zero
int iswupper(wint_t ch); value if ‘ch’ is an uppercase alphabetic character, else returns 0 (Zero).
int islower(int ch); Checks whether 'ch' is a lowercase alphabet. It returns a non-zero value
int iswlower(wint_t ch); if ‘ch’ is a lowercase alphabetic character, else returns 0 (Zero).
int isspace(int ch); Checks whether 'ch' is a white space character. White space characters
int iswspace(wint_t ch); are: ‘ ‘ (Space, ASCII value: 32), ‘\t’ (Horizontal tab, ASCII value: 9), ‘\n’
(Newline, ASCII value: 10), ‘\v’ (Vertical tab, ASCII value: 11), ‘\f’ (Form
feed, ASCII value: 12), ‘\r’ (Carriage return, ASCII value: 13). It returns a
non-zero value if ‘ch’ is a white space character, else returns 0 (Zero).
int isprint(int ch); Checks whether 'ch' is a printable character. Printable characters are the
int iswprint(wint_t ch); ones which have ASCII values between 32 to 126 (included). It returns a
non-zero value if ‘ch’ is a printable character, else returns 0 (Zero).
int iscntrl(int ch); Checks whether 'ch' is a control character. This function returns the
int iswcntrl(wint_t ch); opposite of the function ‘isprint’. A control character is a character that
is not printed on a display. It returns a non-zero value if ‘ch’ is a control
character, else returns 0 (Zero).
int ispunct(int ch); Checks whether 'ch' is a punctuation character. Punctuation characters
int iswpunct(wint_t ch); are all printable characters which are neither alphanumeric nor space. It
returns a non-zero value if ‘ch’ is a punctuation character, else returns 0
(Zero).
int isxdigit(int ch); Checks whether 'ch' is a hexadecimal character. Hexadecimal characters
int iswxdigit(wint_t ch); are: [0-9], [a-f] and [A-F]. It returns a non-zero value if ‘ch’ is a
hexadecimal character, else returns 0 (Zero).
int isgraph(int ch); Checks whether 'ch' has a graphical representation. A character can be
int iswgraph(wint_t ch); graphically represented if it is a printable character except the character
‘ ‘ (Space). It returns a non-zero value if ‘ch’ is a graphical character, else
returns 0 (Zero).

152
 CHARACTER STRING OPERATIONS

9.7.4 STRING CONVERSION FUNCTIONS

What happens if we have a number stored in a string and we need to convert that to a numeral. For
operations like these we have the string conversion functions, which help in converting data to and from
string. These functions are defined in the header file ‘stdlib.h’ and have wide character equivalents.

• ANSI → double atof(const char *str);

Wide Character → double _wtof(const wchar_t *str);
◦ Converts the string 'str' to a double and returns the value. The function discards as many
whitespace characters in front of the string. It then starts converting from the first non
whitespace character and continues until it finds a character which does not fit into a
floating point number. Usage example:
char szValue[] = “3.141592”;
double dblValue = atof(szValue);
wchar_t szValue[] = L“3.141592”;
double dblValue = _wtof(szValue);

• ANSI → int atoi(const char *str);

Wide Character → int _wtoi(const wchar_t *str);
◦ Converts the string 'str' to an integer and returns the value. The function discards as many
whitespace characters in front of the string. It then starts converting from the first non
whitespace character (supports an optional plus or minus sign at the start) and continues
until it finds a character which is not a digit [0-9]. Usage example:
char szValue[] = “12345”;
int iValue = atoi(szValue);
wchar_t szValue[] = L“12345”;
int iValue = _wtoi(szValue);

• ANSI → long int atol(const char *str);

Wide Character → long int _wtol(const wchar_t *str);
◦ Converts the string 'str' to a long integer and returns the value. Usage example:
char szValue[] = “12345”;
long lValue = atol(szValue);
wchar_t szValue[] = L“12345”;
long lValue = _wtol(szValue);

• ANSI → double strtod(const char *str, char **endptr);

Wide Character → double wcstod(const wchar_t *str, wchar_t **endptr);
◦ Converts the string 'str' to a double and returns the value. If ‘endptr’ is not NULL, the
function sets the value of 'endptr' to point to the first character after the number. Usage
example:

153
Quick & Indepth C With Data Structures

char szValues[] = “67.34 84.583”;

char *pEnd = NULL;
double dblValue1 = strtod(szValues, &pEnd);
double dblValue2 = strtod(pEnd, NULL);
wchar_t szValues[] = L“67.34 84.583”;
wchar_t *pEnd = NULL;
double dblValue1 = wcstod(szValues, &pEnd);
double dblValue2 = wcstod(pEnd, NULL);

• ANSI → long int strtol(const char *str, char **endptr, int base);
Wide Character → long int wcstol(const wchar_t *str, wchar_t **endptr, int base);
◦ Converts the string 'str' to a long int and returns the value. If ‘endptr’ is not NULL, the
function sets the value of 'endptr' to point to the first character after the number. If 'base' is
between 2 and 36, it is used as the base of the number. If 'base' is 0 (Zero), the initial
characters of the string pointed to by 'str' are used to determine the base. If the first
character is ‘0’ (Zero) and the second character is not 'x' or 'X', the string is interpreted as
an octal integer. If the first character is ‘0’ (Zero) and the second character is 'x' or 'X', the
string is interpreted as a hexadecimal integer. If the first character is ‘1’ to ‘9’, the string is
interpreted as a decimal integer. Usage example:
char szNumbers[] = "2007 51a1b1 100010010110 0x5A41C8";
char *pEnd = NULL;
long lValue1 = strtol(szNumbers, &pEnd, 10);
long lValue2 = strtol(pEnd, &pEnd, 16);
long lValue3 = strtol(pEnd, &pEnd, 2);
long lValue4 = strtol(pEnd, NULL, 0);
wchar_t szNumbers[] = L"2007 51a1b1 100010010110 0x5A41C8";
wchar_t *pEnd = NULL;
long lValue1 = wcstol(szNumbers, &pEnd, 10);
long lValue2 = wcstol(pEnd, &pEnd, 16);
long lValue3 = wcstol(pEnd, &pEnd, 2);
long lValue4 = wcstol(pEnd, NULL, 0);

• ANSI → unsigned long int strtoul(const char *str, char **endptr, int base);
Wide Character → unsigned long int wcstoul(const wchar_t *str, wchar_t **endptr, int base);
◦ This function is similar to ‘strtol/wcstol’ except for the fact that it interprets an unsigned
long integer and returns the value.

Apart from the above functions, there are few functions which are available from C99 or later. These
functions were created to support higher precision and value. They are:
• ANSI → _int64 _atoi64(const char *str);
Wide Character → _int64 _wtoi64(const wchar_t *str);
◦ Converts the string 'str' to a 64-bit integer and returns the value. Usage example:

154
 CHARACTER STRING OPERATIONS

char szValue[] = “123456789”;

_int64 iValue = _atoi64(szValue);
wchar_t szValue[] = L“123456789”;
_int64 iValue = _wtoi64(szValue);

• ANSI → long long int atoll(const char *str);

Wide Character → long long int _wtoll(const wchar_t *str);
◦ Converts the string 'str' to a long long integer and returns the value. Usage example:
char szValue[] = “1234567890123”;
long long llValue = atoll(szValue);
wchar_t szValue[] = L“1234567890123”;
long long llValue = _wtoll(szValue);

• ANSI → float strtof(const char *str, char **endptr);

Wide Character → float wcstof(const wchar_t *str, wchar_t **endptr);
◦ This function is similar to ‘strtod/wcstod’ except for the fact that it converts the string to a
single precision floating point value and returns that value.

• ANSI → long double strtold(const char *str, char **endptr);

Wide Character → long double wcstold(const wchar_t *str, wchar_t **endptr);
◦ This function is similar to ‘strtod/wcstod’ except for the fact that it converts the string to a
long double value and returns that value.

• ANSI → long long strtoll(const char *str, char **endptr, int base);
Wide Character → long long wcstoll(const wchar_t *str, wchar_t **endptr, int base);
◦ This function is similar to ‘strtol/wcstol’ except for the fact that it converts the string to a
long long integer value and returns that value.

• ANSI → unsigned long long strtoull(const char *str, char **endptr, int base);
Wide Character → unsigned long long wcstoull(const wchar_t *str, wchar_t **endptr, int base);
◦ This function is similar to ‘strtol/wcstol’ except for the fact that it converts the string to an
unsigned long long integer value and returns that value.

Till now we have read about string conversion functions which convert a string to only one type of data,
be it an integer, a float or a double. The conversion produces the data corresponding to only a single
type. They also generate only a single value as their output. What if a string consists of multiple values
corresponding to a single or multiple data type? Is there a function which will enable us to extract
multiple values from a single string at one go? The function ‘sscanf’ is the function which will enable us
to do just that. This function is declared in the header file ‘stdio.h’ and has the following declaration:
ANSI → int sscanf(const char *str, const char *format, ...);
Wide Character → int swscanf(const wchar_t *str, const wchar_t *format, ...);
Just like the ‘scanf’ function, this function too accepts variable number of arguments denoted by ‘...’.

155
Quick & Indepth C With Data Structures

The function reads data from ‘str’ and stores them in the parameter list (called additional arguments)
provided to the function after the parameter ‘format’. We need to provide addresses of the variables in
which we want the formatted output to be stored. The format of the desired generated output needs to
be specified in the ‘format’ string and their data type must match with the data type of the variables
specified in additional arguments section. The number of additional arguments must be at-least the
number of format specifiers specified in ‘format’. This function is very similar to the ‘scanf’ function
which we have already studied (refer section 4.2.2). The only difference is that rather than reading the
input from the standard input device (example: console), this function reads the input from the provided
string ‘str’. Usage example:

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
char szValues1[] = "20 is the age of Matt Jones";
char szValues2[] = "Canada 6217 N 3.141592";
char szValues3[] = "360;1.414213;S;Tutan";
int iResult = 0;
char szOutput[50] = { 0 };
int iOutput = 0;
char chOutput = 0;
float fltOutput = 0;

iResult = sscanf(szValues1, "%50s", szOutput);

iResult = sscanf(szValues1, "%d", &iOutput);
iResult = sscanf(szValues1, "%c", &chOutput);
iResult = sscanf(szValues1, "%f", &fltOutput);

/* Print the values. */

printf("szOutput: %s, iOutput: %d, chOutput: %c, fltOutput: %f\n",
szOutput, iOutput, chOutput, fltOutput);

memset(szOutput, 0, sizeof(szOutput));
fltOutput = 0;
iOutput = chOutput = 0;

/* Now, lets get all values at one go. */

iResult = sscanf(szValues2, "%6s %d %c %f", szOutput, &iOutput, &chOutput, &fltOutput);

/* Print the values. */

printf("szOutput: %s, iOutput: %d, chOutput: %c, fltOutput: %f\n",
szOutput, iOutput, chOutput, fltOutput);

memset(szOutput, 0, sizeof(szOutput));
fltOutput = 0;
iOutput = chOutput = 0;

/* Now, lets get all values again. */

iResult = sscanf(szValues3, "%d;%f;%c;%s", &iOutput, &fltOutput, &chOutput, szOutput);

/* Print the values. */

printf("szOutput: %s, iOutput: %d, chOutput: %c, fltOutput: %f\n",
szOutput, iOutput, chOutput, fltOutput);

return 0;
}

156
 CHARACTER STRING OPERATIONS

The above functions help us to convert a string type to values of other data types. Is there a way to do
the just opposite i.e. convert values of other data types to a string? We can use the ‘sprintf’ function to
do just that. This function too is declared in the header file ‘stdio.h’ and has the following declaration:
ANSI → int sprintf(char *str, const char *format, ...);
Wide Character → int swprintf(wchar_t *str, const wchar_t *format, ...);
Just like the ‘printf’ function, this function too accepts variable number of arguments. The function sets
the string ‘str’ with the values specified in additional arguments. We need to provide the list of
values/variables in the additional argument section. The format specifiers corresponding to the values
in the argument list need to be specified in the ‘format’ string and their data type must match with the
data type of the values/variables. The number of additional arguments must be same as the number of
format specifiers specified in ‘format’. The function composes a string with the same text that would be
printed if 'format' was used on the 'printf' function. The only difference is that rather than producing
the output to the standard output device (example: console), this function generates the output in the
provided string ‘str’. The size of the buffer in 'str' should be large enough to contain the entire resulting
string. A terminating ‘\0’ (NULL) character is automatically appended at the end of the string. Usage
example:

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
char szOutput[80] = { 0 };
int iResult = 0;
char szInput[] = "Gingerbread";
int iInput = 2739;
char chInput = 'K';
float fltInput = 3.141592;

sprintf(szOutput, "%s", szInput);

printf("szOutput: %s\n", szOutput);

sprintf(szOutput, "%d", iInput);

printf("szOutput: %s\n", szOutput);

sprintf(szOutput, "%c", chInput);

printf("szOutput: %s\n", szOutput);

sprintf(szOutput, "%f", fltInput);

printf("szOutput: %s\n", szOutput);

memset(szOutput, 0, sizeof(szOutput));

sprintf(szOutput, "%s, %d, %c, %f", szInput, iInput, chInput, fltInput);

printf("szOutput: %s\n", szOutput);

memset(szOutput, 0, sizeof(szOutput));

sprintf(szOutput, "%s, %d, %c, %f\n", "New York", 112, 65, 1.73205);
printf("szOutput: %s", szOutput);

return 0;
}

157
Quick & Indepth C With Data Structures

9.8 SAMPLE PROGRAMS

1. Write a C program which reverses an existing string.

/* Header Files */
#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization */
char szName[] = "Sarah Taylor";
char *ptr1 = NULL;
char *ptr2 = NULL;
char ch = 0;

/* Move the second pointer to the end of the string. */

for (ptr2 = szName; '\0' != *ptr2; ptr2++);

/* Keep swapping element values, until the pointers cross each other.
When the pointers cross, it will mean we have a reversed string.
‘ptr2’ has moved up to the NULL character. We don’t want the NULL
character to be reversed. So, move one place back before continuing. */
for (ptr1 = szName, ptr2--; ptr1 < ptr2; ptr1++, ptr2--)
{
ch = *ptr2;
*ptr2 = *ptr1;
*ptr1 = ch;
}

/* Print the reversed string. */

printf("The reversed string is %s.\n\n", szName);

return 0;
}

2. Let us write a program which finds the frequency of occurrence of a given character within a
string. The program will do a case in-sensitive search for the character.

/* Header Files */
#include <stdio.h>
#include <ctype.h>

/* Defined Values */
/* Maximum string length including the NULL character. */
#define MAX_STRING_LEN 201

int main(void)
{
/* Variable Declaration & Initialization */
char szString[MAX_STRING_LEN] = { 0 };
int idx = 0;
int iCount = 0;
char chToFind = 0;

/* Accept a string from the user. */

printf("Enter the string: ");

158
 CHARACTER STRING OPERATIONS

gets(szString);

/* Accept the character to find. */

printf("Enter the character to find: ");
scanf("%c", &chToFind);

for (idx = 0; szString[idx] != '\0'; idx++)

{
if (toupper(szString[idx]) == toupper(chToFind))
{
iCount++;
}
}

/* Print the result. */

printf("\nThe character '%c' appears %d times within the string '%s'.\n\n",
chToFind, iCount, szString);

return 0;
}

3. Let us write a program which accepts a string and a character from the user. The program will
remove all occurrence of the given character within the string.

/* Header Files */
#include <stdio.h>

/* Defined Values */
/* Maximum string length including the NULL character. */
#define MAX_STRING_LEN 201

int main(void)
{
/* Variable Declaration & Initialization */
char szString[MAX_STRING_LEN] = { 0 };
int idx1 = 0;
int idx2 = 0;
char chToRemove = 0;

/* Accept a string from the user. */

printf("Enter the string: ");
gets(szString);

/* Accept the character to remove. */

printf("Enter the character to remove: ");
scanf("%c", &chToRemove);

/* Remove the character. */

for (; szString[idx2] != '\0'; idx2++)
{
if (szString[idx2] != chToRemove)
{
szString[idx1++] = szString[idx2];
}
}

/* Set the terminating NULL character. */

szString[idx1] = '\0';

/* Print the result. */

159
Quick & Indepth C With Data Structures

printf("\nThe modified string is %s.\n\n", szString);

return 0;
}

4. In our next program we will accept two wide character strings from the user and concatenate
one string to another. We will do so without using the C library functions.

/* Header Files */
#include <stdio.h>

/* Defined Values */
/* Maximum string length including the NULL character. */
#define MAX_STRING_LEN 201

int main(void)
{
/* Variable Declaration & Initialization */
wchar_t wszString1[MAX_STRING_LEN * 2] = { 0 };
wchar_t wszString2[MAX_STRING_LEN] = { 0 };
int idx1 = 0;
int idx2 = 0;

/* Accept the first string from the user. */

wprintf(L"Enter the first string (no space): ");
wscanf(L"%200s", wszString1);

/* Accept the second string from the user. */

wprintf(L"Enter the second string to concatenate (no space): ");
wscanf(L"%200s", wszString2);

/* Go to the end of the first string. */

for (; wszString1[idx1] != L'\0'; idx1++);

/* Concatenate the strings. */

while (idx1 < (MAX_STRING_LEN * 2 - 1) && wszString2[idx2] != L'\0')
{
wszString1[idx1++] = wszString2[idx2++];
}

wszString1[idx1] = L'\0';

/* Print the result. */

wprintf(L"\nThe concatenated string is %s.\n\n", wszString1);

return 0;
}

5. Let us write a program which will accept a wide character string from the user. The program will
then count the number of vowels, consonants, digits and spaces within the string.

/* Header Files */
#include <stdio.h>
#include <ctype.h>

/* Defined Values */

160
 CHARACTER STRING OPERATIONS

/* Maximum string length including the NULL character. */

#define MAX_STRING_LEN 201

int main(void)
{
/* Variable Declaration & Initialization */
wchar_t wszString[MAX_STRING_LEN] = { 0 };
wchar_t wch = 0;
int idx = 0;
int iVowels = 0;
int iConsonants = 0;
int iDigits = 0;
int iSpaces = 0;

/* Accept the string from the user. */

wprintf(L"Enter the string: ");
fgetws(wszString, MAX_STRING_LEN, stdin);

/* Count the characters. */

for (; wszString[idx] != L'\0'; idx++)
{
wch = towupper(wszString[idx]);
if (L'A' == wch || L'E' == wch || L'I' == wch ||
L'O' == wch || L'U' == wch)
{
iVowels++;
}
else if (wch >= L'A' && wch <= L'Z')
{
iConsonants++;
}
else if (wszString[idx] >= L'0' && wszString[idx] <= L'9')
{
iDigits++;
}
else if (wszString[idx] == L' ')
{
iSpaces++;
}
}

/* Print the results. */

wprintf(L"\nThere are %d vowels, %d consonants, %d digits and %d spaces.\n\n",
iVowels, iConsonants, iDigits, iSpaces);

return 0;
}

161
Quick & Indepth C With Data Structures

STRUCTURES AND UNIONS

[CHAPTER-10]

10.1 INTRODUCTION
We have seen the use of arrays to represent a group of items which belong to the same data type (ex:
char or int). But, is there a way to represent a group of items belonging to different data types using a
single name? C supports such a user-defined data type called a ‘structure’. A structure is a collection of
logically related fields belonging to same or different data type. For example, we may use a structure to
represent the details of each student in a class. The structure may contain the student’s registration
number, name, roll number and marks. We may have the registration number and roll number as of
integer data type, the name as of string (character array), and the marks as integer array. We may have
all these fields declared together within a structure and give the structure a name. So, structures help
us organize multiple logically related fields together in a more meaningful way. Another concept called
‘unions’ is a similar concept to structure but differs in memory/storage usage. We will discuss about
unions a little later in this chapter. For now, let us concentrate on structures.

10.2 DEFINING A STRUCTURE

When we define a structure, we specify the format and composition of the structure on which the
structure variables will be based. The structure variables are variables which are of the structure data
type and contain values as per the structure definition. Assume we are having a grocery shop where we
are using a structure to store the details of each grocery item. We need to store details like item name,
item code, item rate and quantity in stock. We will define the structure as:
struct Item_Details
{
unsigned int uCode;
char szName[100];
float fltRate;
double dblQty;
};

The above structure has the fields ‘uCode’, ‘szName’, ‘fltRate’ and ‘dblQty’. The fields of a structure are
called structure elements or members. Each structure member may be of a different data type and may
also be an array or even of another structure type. ‘Item_Details’ is the name of the structure and is
also called the structure tag. The structure tag is subsequently used in the program to declare variables
belonging to the structure type. We can declare structure variables corresponding to our structure in
the following manner:
struct Item_Details Item1;
struct Item_Details Item2, Item3, Item4;

162
 STRUCTURES AND UNIONS

The general format of defining a structure is as follows:

struct tag_name
{
Data-Type member1;
Data-Type member2;
Data-Type member3;
.
.
};

And the general format for declaring the structure variables is:
struct tag_name Value1, Value2, Value3;

As we can see that while specifying the structure type (also during declaration of the structure
variables), we need to specify the term ‘struct’ followed by its tag name. This may be a bit inconvenient
and we may want an easier way for specifying the structure type. For this, we may type-define the
structure to a name, and use that name in places wherever the structure type needs to be specified.
typedef struct tag_name
{
Data-Type member1;
Data-Type member2;
Data-Type member3;
.
.
}tag_alias;

The type-defined name helps us to avoid typing the term ‘struct’, making our code appear a little
cleaner. It also helps us to achieve the same functionality using a little less keystrokes. But, using the
alias name has some disadvantages too. A type-defined name hides the actual type, i.e. by looking at
the code we cannot understand whether the alias belongs to a structure, an enum or of any other data
type. Once type-defined, we may use the alias name to declare variables belonging to the structure
type.
tag_alias Value1;

We may also define the structure and declare the structure variables at one go. Consider the following
example (Please note that in this case, we are not using the typedef keyword):
struct Item_Details
{
unsigned int uCode;
char szName[100];
float fltRate;
double dblQty;
} Item1, Item2, Item3;

163
Quick & Indepth C With Data Structures

The above statement defines the format of the structure and also declares three variables ‘Item1’,
‘Item2’ and ‘Item3’ belonging to the structure type. We may also define a structure without a tag_name
when we declare the variables corresponding to the structure as in the above case (i.e. during structure
definition itself). We may use the already declared variables in our program, but will not be able to
declare any more variables later on in our program, as the structure does not possess any name. This
kind of structures are called name-less structures and are not recommended to be used.

We can summarize the following related to a structure:

• We must provide a semi-colon at the end of each member declaration.
• The structure definition must also end with a semi-colon.
• The structure definition must be enclosed within parenthesis.
• The data type of each member may be different and may even be an array, union or another
structure type. We may also have pointers as members of a structure.
• We may type-define the structure to have an alias name. We may then use the alias name in
places where the structure type needs to be specified.
• We may declare variables of the structure during the structure definition itself.
• We may create a name-less structure i.e. a structure without a tag-name, when we declare the
structure variables during the structure definition itself.

Every variable of the structure type contain the same format as defined in the structure tag and are the
ones which actually carry the data. The memory allocation of the structure is as per the defined format.
For our structure ‘Item_Details’, the memory space corresponding to the structure variables will have
‘uCode’ appearing first followed by the character array ‘szName’, the float variable ‘fltRate’ and ending
with the double ‘dblQty’. All the members of the structure will appear consecutively in memory and in
the order as specified within the structure definition. Though the members appear consecutively, it will
be wrong to assume that the next member will appear immediately at the next memory location after
the end of the previous member. For performance reasons, some compilers align the structure members
on specific byte boundaries like 4 bytes, 8 bytes, 16 bytes etc. Assume that the members are aligned on
8 bytes boundaries and our first member ‘uCode’ starts from the memory location 96 and takes 4 bytes
of memory. Our next member ‘szName’ will start from the memory location 104 (not 100) as the
members are aligned on 8 bytes boundaries. We do not need to bother about byte boundaries as it is
automatically managed by the compiler and we access the structure members using their names.

10.3 ACCESSING MEMBERS

The members of a structure can be accessed using the dot operator ‘.’ which is also called the member
operator or the period operator. The usage of a structure member is very similar to a simple variable of
the same type, except that it needs to be used in conjunction with the structure variable and separated
from the structure variable by the dot operator.
Item1.uCode
Item1.fltRate
Item2.szName

164
 STRUCTURES AND UNIONS

We can not only access the members of the structure but also assign values to them in the same
manner as we do with other variables of the same data type. Consider the following statements:
struct Item_Details Item1;
Item1.uCode = 1;
Item1.dblQty = 106.50;
strcpy(Item1.szName, “Vanilla ice cream”);
scanf(“%f”, &Item1.fltRate);
printf(“Total amount of %s is %lf.”, Item1.szName, Item1.dblQty * Item1.fltRate);

Let us understand the discussed concepts using a simple program. Here we will accept the item details
within a loop and continue doing so till the user wants to. We will then print the accepted values.

/* Header Files */
#include <stdio.h>
#include <ctype.h>

/* Defined Values */
#define ARRAYLEN(x) sizeof(x)/sizeof(x[0])

/* Structure Definition */
struct Item_Details
{
unsigned int uCode;
char szName[100];
float fltRate;
double dblQty;
};

int main(void)
{
/* Variable Declaration & Initialization */
struct Item_Details Item;
size_t cchLen = 0;
char chContinue = 'y';

while ('Y' == toupper(chContinue))

{
/* Initialize the structure variable with 0s. */
memset(&Item, 0, sizeof(Item));

/* Accept the item details. */

printf("Enter the item code: ");
scanf("%u", &Item.uCode);

/* When we use 'fgets' function with 'stdin' as input stream, it

accepts the string from the standard input device ex. Console.
We will read more about 'fgets' function in our chapter on
file management. Here we are avoiding 'scanf' function as it
does not accept spaces within an input string. Using of 'gets'
function should be avoided as it does not check for buffer overflows,
and will write past the buffer if provided with a long string.*/
printf("Enter the item name: ");
fgets(Item.szName, ARRAYLEN(Item.szName), stdin);

/* Get rid of the new line character as 'fgets' function retains

the new line character in the input string. */

165
Quick & Indepth C With Data Structures

cchLen = strlen(Item.szName);
if (0 < cchLen && '\n' == Item.szName[cchLen – 1])
{
Item.szName[cchLen - 1] = '\0';
}

printf("Enter the item rate: ");

scanf("%f", &Item.fltRate);

printf("Enter the item quantity: ");

scanf("%lf", &Item.dblQty);

/* Print the accepted values. */

printf("\n\nThe item details are:");
printf("\nItem code: %u", Item.uCode);
printf("\nItem name: %s", Item.szName);
printf("\nItem rate: %.02f", Item.fltRate);
printf("\nItem quantity: %.02lf", Item.dblQty);
printf("\nTotal amount: %.02lf\n\n", Item.dblQty * Item.fltRate);

/* Check if user wants to continue. */

printf("Continue ('Y' to continue)? ");
chContinue = getchar();

printf("\n\n");
}

return 0;
}

10.4 STRUCTURE INITIALIZATION

A structure is initialized a little differently than other variables. Unlike simple variables, initializing a
structure involves initializing all its member variables. Thus, initialization of structure has a different
format than other variables. Structure variables can be initialized during structure definition or during
variable declaration. Have a look at some examples:

struct Item_Details
{
unsigned int uCode;
char szName[100];
float fltRate;
double dblQty;
} Item1 = { 10, "Dark Chocolate", 5.00, 50.00 };

In the example, 10 gets assigned to ‘uCode’, the string “Dark Chocolate” gets assigned to ‘szName’,
5.00 gets assigned to ‘fltRate’ and 50.00 gets assigned to ‘dblQty’. All these values are assigned to the
members of the structure variable ‘Item1’. There is a one-to-one mapping between the values specified
within the curly braces and the structure members. The order of the values should match the order of
the structure members to which they are to be assigned. Structure variables can also be initialized as:
struct Item_Details Item2 = { 11, “Brown bread”, 1.50, 20.00 };
static struct Item_Details Item3 = { 12, “Darjeeling Tea”, 12.50, 20.50 };

166
 STRUCTURES AND UNIONS

We may also initialize one structure variable from another structure variable in the same way we do with
other variables.

struct Item_Details Item2 = { 11, “Brown bread”, 1.50, 20.00 };

struct Item_Details Item3 = Item2;
static struct Item_Details Item4 = Item3;

As it is evident from the above examples, when one structure variable is assigned to another structure
variable a byte-by-byte copy occurs from the variable on the right to the variable on the left. So, all
members of the left-side variable get the same values as the members of the right-side variable. In the
above example, when Item2 gets assigned to Item3, the members of Item3 also get the same values as
the Item2 members. So, ‘uCode’ gets the value 11, ‘szName’ gets the string “Brown bread”, ‘fltRate’
gets the value 1.50 and ‘dblQty’ gets the value 20.00.

If we want to initialize all the members of a structure variable to a same value, we may do so as below:
struct Item_Details Item2 = { 0 };

The above statement initializes all members of the structure variable to the value 0.

10.5 ARRAY OF STRUCTURE

We can declare array of structures in the same way as we do for other data types. Let us understand
this using a simple program. In this program, we declare an array of structures consisting of 10
elements. Each structure element accepts and holds the marks details of a single student of a class. It
then prints the marks and also the average marks received by each of the students.

#include <stdio.h>

/* Defined Values */
#define MAX_STUDENTS 10

/* Structure Definition */
struct Student_Marks
{
int iMathematics;
int iChemistry;
int iPhysics;
int iBiology;
};

int main(void)
{
/* Variable Declaration & Initialization */
struct Student_Marks rgMarks[MAX_STUDENTS] = { 0 };
int idx = 0;
int iAggregate = 0;

for (; idx < MAX_STUDENTS; idx++)

{
printf("Student #%02d:\n============\n", idx + 1);
printf("Enter the marks obtained in Mathematics: ");

167
Quick & Indepth C With Data Structures

scanf("%d", &rgMarks[idx].iMathematics);

printf("Enter the marks obtained in Chemistry: ");

scanf("%d", &rgMarks[idx].iChemistry);

printf("Enter the marks obtained in Physics: ");

scanf("%d", &rgMarks[idx].iPhysics);

printf("Enter the marks obtained in Biology: ");

scanf("%d", &rgMarks[idx].iBiology);
printf("\n");
}

printf("\n\n");

/* We will first print the headings and then the marks. */

printf("\n\t\t\t\tSTUDENT MARKS\n\t\t\t\t*************\n\t\t");

/* Make each subject column fixed width (11 chars) & left aligned. A tab is also
inserted between two columns to provide proper separation between columns. */
printf("Mathematics\tChemistry \tPhysics \tBiology \tAverage\n");

/* Now, print the marks and the percentage obtained by the students. */
for (idx = 0; idx < MAX_STUDENTS; idx++)
{
printf("\nStudent #%02d:\t", idx + 1);

printf("%02d \t", rgMarks[idx].iMathematics);

printf("%02d \t", rgMarks[idx].iChemistry);
printf("%02d \t", rgMarks[idx].iPhysics);
printf("%02d \t", rgMarks[idx].iBiology);

/* Calculate the aggregate of all the subjects for the student. */

iAggregate = rgMarks[idx].iMathematics + rgMarks[idx].iChemistry +
rgMarks[idx].iPhysics + rgMarks[idx].iBiology;

/* Print the percentage obtained by this student. */

printf("%02.02lf", (double) iAggregate / (double) 4);
}

printf("\n\n");

return 0;
}

Sample output:

STUDENT MARKS
*************
Mathematics Chemistry Physics Biology Average

Student #01: 76 68 72 70 71.50

Student #02: 67 62 69 70 67.00
Student #03: 84 87 82 84 84.25
Student #04: 77 81 78 75 77.75
Student #05: 91 85 81 84 85.25
Student #06: 58 66 63 67 63.50
Student #07: 75 72 77 73 74.25
Student #08: 61 64 69 59 63.25
Student #09: 73 71 69 68 70.25
Student #10: 74 86 72 78 77.50

168
 STRUCTURES AND UNIONS

In the above program we have declared an array ‘rgMarks’ of the structure ‘Student_Marks’. The array
has 10 elements and each element holds the marks details corresponding to a single student. An
element of the array of a structure is accessed in the same way as any other array element.
rgMarks[4] The 4th element of the array.
rgMarks[4].iPhysics The ‘iPhysics’ member of the 4th element of the array.

10.6 POINTER TO STRUCTURE

If we have a pointer to a structure variable, the method of accessing its member element is a bit
different. When used with a pointer variable, a member is accessed using the arrow operator (–>). The
arrow operator is formed using a minus sign followed by a greater than sign. Consider the below
example which implements the above program (only the main() function shown) using a pointer.

int main(void)
{
/* Variable Declaration & Initialization */
struct Student_Marks rgMarks[MAX_STUDENTS] = { 0 };
struct Student_Marks *pMarks = NULL;
int iAggregate = 0;

for (pMarks = rgMarks; (pMarks - rgMarks) < MAX_STUDENTS; pMarks++)

{
printf("Student #%02d:\n============\n", (pMarks - rgMarks) + 1);
printf("Enter the marks obtained in Mathematics: ");
scanf("%d", &pMarks->iMathematics);

printf("Enter the marks obtained in Chemistry: ");

scanf("%d", &pMarks->iChemistry);

printf("Enter the marks obtained in Physics: ");

scanf("%d", &pMarks->iPhysics);

printf("Enter the marks obtained in Biology: ");

scanf("%d", &pMarks->iBiology);
printf("\n");
}

printf("\n\n");

/* We will first print the headings and then the marks. */

printf("\n\t\t\t\tSTUDENT MARKS\n\t\t\t\t*************\n\t\t");

/* Make each subject column fixed width (11 chars) & left aligned. A tab is also
inserted between two columns to provide proper separation between columns. */
printf("Mathematics\tChemistry \tPhysics \tBiology \tAverage\n");

/* Now, print the marks and the percentage obtained by the students. */
for (pMarks = rgMarks; (pMarks - rgMarks) < MAX_STUDENTS; pMarks++)
{
printf("\nStudent #%02d:\t", (pMarks - rgMarks) + 1);

printf("%02d \t", pMarks->iMathematics);

printf("%02d \t", pMarks->iChemistry);
printf("%02d \t", pMarks->iPhysics);
printf("%02d \t", pMarks->iBiology);

169
Quick & Indepth C With Data Structures

/* Calculate the aggregate of all the subjects for the student. */

iAggregate = pMarks->iMathematics + pMarks->iChemistry +
pMarks->iPhysics + pMarks->iBiology;

/* Print the percentage obtained by this student. */

printf("%02.02lf", (double) iAggregate / (double) 4);
}

printf("\n\n");

return 0;
}

In the above program we have declared a pointer ‘pMarks’ of the structure ‘Student_Marks’. The
statement ‘pMarks = rgMarks’ assigns the address of the 0th element of the array to the pointer
‘pMarks’. This statement is equivalent to using the statement ‘pMarks = &rgMarks[0]’. As per pointer
arithmetic, incrementing ‘pMarks’, makes the pointer point to the next element of the array. Similarly,
decrementing its value makes it point to the previous element. Subtracting one pointer from another
gives the number of elements between the two pointers. So the expression ‘pMarks – rgMarks’ gives the
number of elements between the element pointed by ‘pMarks’ and start of the array. A member of an
element of the array is accessed as below:
pMarks->iMathematics
pMarks->iChemistry
pMarks->iPhysics
pMarks->iBiology

We can also use the following syntax to access a member of the element pointed by the pointer.
(*pMarks).iMathematics
(*pMarks).iChemistry
(*pMarks).iPhysics
(*pMarks).iBiology

The above four statements initially dereference the pointer ‘(*pMarks)’ and then use the dot operator to
access the structure member. The parentheses around ‘*pMarks’ is important as the dot operator has
higher precedence than the pointer indirection operator (*).

While using the operators ‘->’ and ‘.’, we must be careful regarding their precedence. They have one of
the highest precedence of all the operators (refer section 3.12). Consider few examples:
(++pMarks)->iPhysics Increments ‘pMarks’ first and then accesses the value of ‘iPhysics’.
++pMarks->iPhysics Increments ‘iPhysics’ corresponding to the element pointed by ‘pMarks’.
pMarks++->iPhysics Accesses ‘iPhysics’ and then increments ‘pMarks’.

10.7 NESTED STRUCTURES

When a structure variable is declared within another structure, it is called nesting of structures. Nested
structures can be of two types:

170
 STRUCTURES AND UNIONS

1. Externally defined nested structures: When a structure variable is declared (as a member)
within another structure but, the definition of the former structure is outside the latter. Some
examples of such nested structures are:
struct Struct1
{
int iValue1;
};

struct Struct2
{
struct Struct1 firstStruct;
int iValue2;
};

In our example, the structure ‘Struct1’ is defined outside ‘Struct2’ but, is declared as a member
of ‘Struct2’ using the variable ‘firstStruct’. If we declare a variable ‘secondStruct’ of type
‘Struct2’, we may access the ‘Struct1’ member ‘iValue1’ as ‘secondStruct.firstStruct.iValue1’.

2. Embedded nested structures: When one structure is defined within another structure, it is
called an embedded nested structure. Have a look at the below example:
struct Struct1
{
struct Struct2
{
int iValue2;
};
int iValue1;
};

The structure ‘Struct2’ is defined within ‘Struct1’ which makes it an embedded nested structure.
If we declare a variable ‘firstStruct’ of type ‘Struct1’, we may access the ‘Struct2’ member
‘iValue2’ as ‘firstStruct.iValue2’. We can access ‘iValue2’ directly as a member of ‘firstStruct’
because ‘Struct2’ is embedded nested and no structure variable of ‘Struct2’ is declared within
‘Struct1’. Consider another example:
struct Struct1
{
struct Struct2
{
int iValue2;
}secondStruct;

int iValue1;
};

171
Quick & Indepth C With Data Structures

In the previous example, if we declare a variable ‘firstStruct’ of type ‘Struct1’, we may access
‘Struct2’ member ‘iValue2’ as ‘firstStruct.secondStruct.iValue2’. This is because we have now
declared a variable ‘secondStruct’ of type ‘Struct2’. This variable is declared within ‘Struct1’.

Let us now write a program which accepts the student details of a class. The program will define a
structure called ‘Student_Profile’ which contains the profile (registration number, name and address) of
a student. The program will define another structure called ‘Student_Marks’ which will hold the marks
obtained by each student of the class. Both ‘Student_Profile’ and ‘Student_Marks’ will be nested within
another structure ‘Student_Details’.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_STUDENTS 10

/* Structure Definitions */
struct Student_Profile
{
int iRegistration;
char szName[50];
};

struct Student_Details
{
int iRoll;

struct Student_Profile Profile;

struct Student_Marks
{
int iMathematics;
int iChemistry;
int iPhysics;
int iBiology;
}Marks;
};

int main(void)
{
/* Variable Declaration & Initialization */
struct Student_Details rgDetails[MAX_STUDENTS] = { 0 };
int idx = 0;

for (; idx < MAX_STUDENTS; idx++)

{
printf("Student #%02d:\n============\n", idx + 1);
printf("Enter the student's registration number: ");
scanf("%d", &rgDetails[idx].Profile.iRegistration);

printf("Enter the student's name: ");

gets(rgDetails[idx].Profile.szName);

printf("Enter the student's roll number: ");

scanf("%d", &rgDetails[idx].iRoll);

printf("Enter the marks obtained in Mathematics: ");

172
 STRUCTURES AND UNIONS

scanf("%d", &rgDetails[idx].Marks.iMathematics);

printf("Enter the marks obtained in Chemistry: ");

scanf("%d", &rgDetails[idx].Marks.iChemistry);

printf("Enter the marks obtained in Physics: ");

scanf("%d", &rgDetails[idx].Marks.iPhysics);

printf("Enter the marks obtained in Biology: ");

scanf("%d", &rgDetails[idx].Marks.iBiology);
printf("\n");
}

printf("\n\n");

/* We will first print the headings and then the marks. */

printf("\n\t\t\tSTUDENT MARKS\n\t\t\t*************\n");

/* Now, print the marks by the students. */

for (idx = 0; idx < MAX_STUDENTS; idx++)
{
printf("\n\nRegistration: %d", rgDetails[idx].Profile.iRegistration);
printf("\nName : %s\t", rgDetails[idx].Profile.szName);

printf("\nRoll Number : %03d", rgDetails[idx].iRoll);

printf("\nMathematics : %02d", rgDetails[idx].Marks.iMathematics);

printf("\nChemistry : %02d", rgDetails[idx].Marks.iChemistry);
printf("\nPhysics : %02d", rgDetails[idx].Marks.iPhysics);
printf("\nBiology : %02d", rgDetails[idx].Marks.iBiology);
}

printf("\n\n");

return 0;
}

10.8 BIT FIELDS

A ‘char’ data type takes 1 byte, a ‘short’ takes 2 bytes and an ‘int’ takes 2, 4 or 8 bytes depending on
the compiler. In situations where we do not require an entire byte for a structure member, we may use a
bit field. This saves memory space and we can accommodate more data by using less memory. The
general form of bit field definition is:
struct tag-name
{
Data-Type field1: bit-length;
Data-Type field2: bit-length;
….
….
Data-Type fieldN: bit-length;
};

The size of a bit field can vary from 1 bit to the total bit-length of the data type if it had no bit-field
specified. So, a ‘short int’ data type can have bit-field length of 1 bit to a maximum of 16 bits which is
the size of a ‘short int’ data type. A signed bit field should have at-least 2 bits (1 extra bit for the sign).

173
Quick & Indepth C With Data Structures

The maximum value that can be stored in a bit field depends on the bit-field length. If the bit-field
length is specified as ‘n’, the maximum value that can be stored in the field is 2 n-1.
The internal representation of the bits vary depending on the machine. They may be stored from left to
right, or from right to left depending on the machine architecture.

The points worth noting regarding bit-fields are:

• The first bit-field starts from a word boundary.
• The address of a bit-field member cannot be accessed as they may not start at a byte boundary.
This means that we cannot do pointer operations on a bit-field or use the ‘scanf’ function to
read value into a bit-field.
• A special unnamed bit field of size 0 (zero) is used to force alignment on to next boundary. This
makes the next member always starts from a byte boundary.
• We cannot have an array of bit fields.
• The bit-field length of a field cannot be more than the number of bits in the specified data-type.
It has to have a minimum length of 1. Signed bit-fields must have a minimum length of 2.
• We should be careful in assigning values in the range which a bit-field can hold. The behavior
due to assignment of out of range values is compiler specific.
• Consecutive bit-fields of same data-type share the bits and though the sum of all the bits may
not be an exact multiple of the data-type size (ex: 8 bits for char, 16 bits for short int), all these
fields together will occupy a size equal to a multiple of the data-type size in bytes.

In our next program we will pick the current date-time from the system. We use the C function 'time' to
get the current time. The function accepts a pointer to a variable of type 'time_t'. 'time_t' is generally a
type-defined name of any integer data type. The function populates this variable with a value which is
equal to the number of seconds elapsed since 00:00 hours, Jan 1, 1970 UTC. The value returned in the
'time_t' variable is passed to another C function called 'localtime'. This function in turn returns a 'tm'
structure pointer. This returned structure variable has a global scope which allows its value to be
retained even after the function has ended. The variable's value can be changed in any subsequent call
to functions ‘gmtime’ or ‘localtime’. ‘gmtime’ is a similar function like ‘localtime’ except that it returns
the time in UTC. The 'tm' structure has the following members:
Member Data Type Description Range
tm_sec int Seconds passed after the last minute 0-59
tm_min int Minutes passed after the last hour 0-59
tm_hour int Hours passed since midnight 0-23
tm_mday int Day of the month 1-31
tm_mon int Months since January 0-11
tm_year int Years since 1900
tm_wday int Days since Sunday i.e. weekday 0-6
tm_yday int Days since January 1 of the year 0-365
tm_isdst int Daylight Saving Time flag. It is greater than 0 (zero) if Daylight
Saving Time is in effect, 0 (zero) if Daylight Saving Time is not in
effect, and less than 0 (zero) if the information is unavailable.

174
 STRUCTURES AND UNIONS

After getting the current time, we assign the values to our structure with the bit-fields. Do notice, in our
structure we have defined an unnamed field with 0 (zero) bits. This field forces the alignment on to the
next boundary. So, the time related field (uHour) starts from the next boundary. So, all the date related
fields (uYear, uMonth and uDate) together occupy starting 21 bits (of unsigned int) leaving rest of the
bits unused. If the 'unsigned int' data type occupies 32 bits (4 bytes) of memory, the rest 11 bits remain
unused. If 'unsigned int' occupies 16 bits (2 bytes) of memory, 2 consecutive unsigned int locations
(totalling 4 bytes) will be used. The first 'unsigned int' will be fully used and 5 bits from the next one will
be used. The rest 11 bits from the second 'unsigned int' will remain unused. The time field (uHour) will
start from a fresh byte boundary.
After assigning the values to our structure variable, we print the values. In the 'printf' function, we use
the format specifier of the data type of the bit-fields for printing their values.

/* Header Files */
#include <stdio.h>
#include <time.h>

/* Structure Definitions */
struct DateTime
{
unsigned uYear : 12;
unsigned uMonth : 4;
unsigned uDate : 5;
unsigned : 0;
unsigned uHour : 5;
unsigned uMinute : 6;
unsigned uSecond : 6;
};

int main(void)
{
/* Variable Declaration & Initialization */
struct DateTime datetime = { 0 };
struct tm *pCurTime = { 0 };
time_t itime = 0;

time(&itime);
pCurTime = localtime(&itime);

if (pCurTime)
{
datetime.uYear = pCurTime->tm_year + 1900;
datetime.uMonth = pCurTime->tm_mon + 1;
datetime.uDate = pCurTime->tm_mday;
datetime.uHour = pCurTime->tm_hour;
datetime.uMinute = pCurTime->tm_min;
datetime.uSecond = pCurTime->tm_sec;
}

printf("Size: %u\n\n", sizeof(datetime));

printf("Todays date : %u/%u/%u\n", datetime.uYear, datetime.uMonth,

datetime.uDate);
printf("Current time: %u:%u:%u\n\n", datetime.uHour, datetime.uMinute,
datetime.uSecond);

return 0;
}

175
Quick & Indepth C With Data Structures

10.9 UNIONS
Unions are a concept borrowed from structures. As they are very similar to structures, their syntax is
also similar to structures. The way a union’s member is accessed, its initialization, arrays of unions,
pointers to unions etc. are all very similar to structures. The major distinction between structures and
unions is in terms of memory usage. In structures, each member has its own separate storage location,
whereas all the members of a union are assigned the same memory location. This implies that although
a union may have multiple members of different data types, we may use (store and read) only one
member at a time. The memory used by a union is equal to the size of its biggest member i.e. the
member which occupies the highest amount of storage. The general form of a union is as follows:
union tag_name
{
Data-Type member1;
Data-Type member2;
Data-Type member3;
.
.
};

We may type-define union in the same way we do with structures. We may also declare union variables
in the same way we do for structure variables. Consider the below example of a union:
union Values
{
char chVal;
short iVal;
double dblVal;
char szName[10];
};

In our above example, the union ‘Values’ has four members. We may use only one member at a time as
all the members occupy the same memory location. The size of the union is equal to its largest member
which in our case is ‘szName’ occupying 10 bytes of storage. So, the storage allocation of all the
members of the union will be:

We can use the same syntax as that of structures while accessing union members. For example, if we
declare a variable ‘MyValues’ of the above union type, we may access its members as:

176
 STRUCTURES AND UNIONS

union Values MyValues;

MyValues.chVal = ‘A’;
strcpy(MyValues.szName, “Jones”);

While accessing a member, we should make sure that we are accessing the member whose value is
currently stored. Union allocates a memory space which can be used by any one of its members at a
time. When a different member is assigned a value, the new value supersedes the value stored in the
previous member. The following group of statements will produce erroneous output as we are accessing
a member which is different from the one which currently contains the value.
union Values MyValues;
MyValues.iVal = 793;
strcpy(MyValues.szName, “Hanna”);
printf(“Value is %d.”, MyValues.iVal);

A union may be nested inside another structure or union. A structure may also be nested inside a union.

In our next program we will calculate the perimeter of different geometrical shapes. We will define a
union which will hold the dimensions needed to calculate the perimeter of the shape.

#include <stdio.h>

/* Structure and Union Definitions */

union Dimensions
{
struct Rectangle
{
int x;
int y;

}rect;

struct Square
{
int x;

}square;

struct Circle
{
float fltRadius;

}circle;
};

int main(void)
{
union Dimensions dim;
int iOption = 0;
const double PI = 3.141592;

while (4 != iOption)
{
printf("\n1. Rectangle\n2. Square\n3. Circle\n4. Exit\n\n");

177
Quick & Indepth C With Data Structures

iOption = 0;
while (iOption < 1 || iOption > 4)
{
printf("Enter the option: ");
scanf("%d", &iOption);
}

if (4 == iOption)
{
break;
}

switch (iOption)
{
case 1:
printf("\n\nEnter the length of the rectangle: ");
scanf("%d", &dim.rect.x);
printf("Enter the width of the rectangle: ");
scanf("%d", &dim.rect.y);

printf("\nThe perimeter of the rectangle is: %d.\n",

2 * (dim.rect.x + dim.rect.y));

break;

case 2:
printf("\n\nEnter the length of the sides of the square: ");
scanf("%d", &dim.square.x);

printf("\nThe perimeter of the square is: %d.\n", 4 * dim.square.x);

break;

case 3:
printf("\n\nEnter the radius of the circle: ");
scanf("%f", &dim.circle.fltRadius);

printf("\nThe perimeter of the circle is: %lf.\n",

(2 * PI * dim.circle.fltRadius));

break;
}
}

return 0;
}

178
 USER DEFINED FUNCTIONS

USER DEFINED FUNCTIONS

[CHAPTER-11]

11.1 INTRODUCTION
In our previous chapters, we have used many system functions which help us achieve certain specific
functionalities. We also defined the ‘main’ function in all our programs which forms the entry point of
the program. As discussed earlier, a function is a group of statements which perform certain specific
tasks. A C program is divided into functions and a function is called from other functions to make it
perform the required tasks. Dividing a program into multiple functions has the following advantages:
1. Functions improve program readability and understanding: It is much easier to understand a
program if it is broken down into multiple pieces and the functionality of each piece is
understood separately. It also improves the readability as we may be looking at functions
entirely contained within a single screen in the IDE, rather than functions which require a lot of
scrolling.
2. Increase in code flexibility: As a function is generally short and does specific tasks, making
changes to it becomes easy till its input, output and basic functionality remains the same. Also,
there is less chance of regression when fixing bugs within a function.
3. Increase in code re-usability: Shifting a specific task to a function helps us to avoid writing the
same piece of code again and again. If a bug is found in a group of statements repeated across
multiple locations, we will require to make the fixes across all such locations. If the repeatative
code is shifted to a function, we will then need to make the fix only within the function. Writing
the same code again and again increases program complexity and must be avoided.
4. Functions achieve procedural abstraction: Once a function is written and tested, it serves as a
black box for the caller functions. All that a programmer (when writing the caller functions)
needs to know for invoking a function is its name and the parameters that it expects.
5. Simplifies program development: When working on a big project, a program may be divided
between multiple programmers with each programmer assigned to implement a specific
functionality using a group of functions. This enables us to implement the project in parallel.
6. Program testing becomes easy: Each function may be tested separately without the need for
the entire program to be ready for testing. The function may be provided dummy inputs and the
actual output verified against expected output. Also when a bug is found, we need to
concentrate only on the faulty function.
7. Code sharing becomes easy: A function written for one program may be used by other
programs too. This is another form of code re-usability.

C functions may be classified into two categories viz. Library functions (refer section 1.2) and user-
defined functions. ‘printf’ and ‘scanf’ are examples of library functions and ‘main’ is an example of
user-defined function. The main distinction between these two categories is that library functions come
bundled with the compiler and do not require to be written by the programmer, whereas user-defined
functions have to be developed by the programmer at the time of writing the program.

179
Quick & Indepth C With Data Structures

Before writing a program, we should always create a design layout of the entire program by dividing it
into functions with proper understanding regarding a function’s task, the kind of output it will generate
and the list of inputs it will accept. We should try to move any functionality which may appear in
multiple places to functions.

An example of a program structure broken up into multiple modules using functions is shown below:

In the above figure, the function ‘main’ calls the four functions ‘function1’, ‘function2’, ‘function3’ and
‘function4’ for getting its work done. ‘function1’ in turn calls the two functions ‘function5’ and
‘function6’. ‘function2’ calls the functions ‘function6’ and ‘function7’ to get some of its tasks get done by
them. Similarly, ‘function3’ calls the functions ‘function8’, ‘function9’ and ‘function10’ to get some of its
work done. The function ‘function4’ too calls the function ‘function10’. We can see that the function
‘function6’ gets called by two functions ‘function1’ and ‘function2’ respectively to help them perform
their work. Similarly, ‘function10’ gets called by the two functions ‘function3’ and ‘function4’.

11.2 THE STRUCTURE OF A C FUNCTION

Every C function have the following form:
Return-Data-Type function-Name(Comma separated argument list)
{
Local Variable Declarations;

Executable statement 1;
Executable statement 2;
.…..
.…..

return value/expression;
}

Some of the parts of a function are optional and may be skipped.

11.2.1 FUNCTION NAME

Every function is recognized by a name. Naming a function is similar to naming any other C identifier
(refer section 1.6). A function name must be unique across the program and is case-sensitive. We

180
 USER DEFINED FUNCTIONS

should be careful to avoid naming a function with the same name as a library function.

11.2.2 FUNCTION RETURN TYPE AND VALUES

How does a function let its caller know the result of its operation? The called function can return only
one value to its caller indicating the result of its operation. If it has multiple values to return, it can do
so using pointers as arguments to the function. It can also return multiple values to the caller by
enclosing all of them within a structure variable and returning it. A function can return a value using the
return statement. The return statement can be written as any one of the below:
• return;
• return (value/expression);
The first ‘return’ statement does not return a value and is simply used to return the control back to the
calling function. When a ‘return’ is encountered, all successive statements within the function is ignored
and the control is immediately transferred to the calling function. As for the second ‘return’ statement,
it immediately returns the value or the result of the expression following the ‘return’ statement to the
caller. The type of value returned by the ‘return’ statement must match the return data type specified in
the function declaration (before the function name).
int sum(int x, int y) long Sum(int x, int y) int CheckValue(int x)
{ { {
return (x + y); long lSum = (long) x + (long) y; if (x <= 0)
} return lSum; return 0;
} else
return 1;
}

A function may return nothing, in which case ‘void’ is specified as the return type of the function. In
such a case, we need not specify any ‘return’ statement for returning control from the called function to
the caller. If any premature return of control is needed (when we do not want to execute
successive/later statements based on a condition check), we may use the first ‘return’ statement (i.e.
the ‘return’ statement without any accompanying value/expression).
void IsZero(int x)
{
if (x == 0)
{
printf(“x is zero”);
return;
}
printf(“x is not zero”);
}

When we return a variable from a function, we actually return the value contained within the variable.
Similarly when we return a pointer from a function, we are actually returning the memory address held
in the pointer variable and not the variable itself. Care must be taken so that we do not return address
of any local variable (variables declared within a function) of the called function, because when a

181
Quick & Indepth C With Data Structures

function returns, all local variables of the function cease to exist (except static or dynamically allocated
variables). So, if the caller function tries to access the returned memory address, it will cause a memory
violation error.

11.2.3 FUNCTION ARGUMENTS

The argument variables or parameter variables receive values from the calling function, thus providing a
means to pass data from the calling function to the called function. The argument or parameter list
contains a list of variable declarations separated by commas.

An argument or parameter is the data which is taken as input by a function. There are two types of
arguments or parameters. The arguments which are passed in the function call are known as actual
arguments or actual parameters. The arguments which are received by the called function are known
as formal arguments or formal parameters.
int sum(int argX, int argY)
{
return (argX + argY);
}

int main(void)
{
int x = 36, y = 64;
sum(x, y);
return 0;
}

In the above program, the ‘x’ and ‘y’ specified within the call to the function ‘sum’ are the actual
arguments (parameters). The variables ‘argX’ and ‘argY’ specified within the function ‘sum’ are the
formal arguments (parameters). There is a one to one mapping between the actual arguments and the
formal arguments. The values contained in the actual arguments are assigned to the formal arguments
when the control comes within the called function.

The argument list of a function is also optional. In such a case, we may specify ‘void’ as the argument
or leave it empty.

If the formal argument list is left empty, the function is assumed to accept an unspecified number of
arguments. So, all calls to the function become valid (called with or without any arguments). Remember,
the parentheses (first brackets) around the argument list is required even if there are no arguments
present.

If ‘void’ is specified as the formal argument (refer to the main function in the last example), the function
is believed to be taking no arguments. So if you call this function with any arguments, the compiler will
raise an error.

182
 USER DEFINED FUNCTIONS

Based on the arguments and return values, functions can be categorized into four types. They are:
• Function without arguments and without return value.
• Function without arguments but with return value.
• Function with arguments and also with return value.
• Function with arguments but without return value.

11.2.4 FUNCTION BODY

The body of a function is enclosed within the opening and closing curly braces. If we are working with
ANSI C, the local function level variables are declared at the start of a function just after the opening
curly braces for the function. In case of C99 or later, we can declare the variables anywhere in the
function before their use, though it is a good practice to declare the variables at the start of the
function to improve readability.

All variables declared within the function exist only till the control within the function remains. As soon
as a function returns, these variables lose their existence. These variables will be re-declared in the next
call to the function. There are few exceptions. A variable declared as ‘static’ remains into existence and
retains its value even when the function exits. This variable will not be re-declared (or re-initialized) and
will continue with its previous value during the next call to the function. Dynamically allocated memory
space (if not freed) also remains into existence even when the function exits. We will read more about
them in our chapter on ‘Dynamic Memory Allocation’. Though any pointer variable (if not declared as
static) pointing to such dynamically allocated memory ceases to exist, the allocated memory space if
not freed, stays. If we return the address of the allocated memory to the caller function, the caller may
still access this memory location once the called function has ended.

11.2.5 EXAMPLE PROGRAMS

In our first multi-function program, we will keep it simple. We will declare two variables 'x' and 'y' in
'main', initialize them and then pass them as actual parameters to another function 'PrintValues'.
'PrintValues' declares two integer variables 'ParamX' and 'ParamY' as its formal parameters. There is a
one to one mapping between the actual parameters and the formal parameters. So, the values of the
variables 'x' and 'y' are assigned to the formal parameters 'ParamX' and 'ParamY' within 'PrintValues'.
'PrintValues' gets the values in the variables 'ParamX' and 'ParamY' and prints their values.

#include <stdio.h>

void PrintValues(int ParamX,

int ParamY)
{
printf("\nThe two numbers are %d and %d.", ParamX, ParamY);
}

int main(void)
{
/* Variable Declaration & Initialization */
int x = 50;
int y = 500;

183
Quick & Indepth C With Data Structures

PrintValues(x, y);

return 0;
}

The function ‘PrintValues’ is an example of a function accepting arguments but having no return values.

In our next program, we will pass two values 44 and 55 from ‘main’ to another function ‘sum’. The
function ‘sum’ will add the two numbers and return the sum of the numbers to the caller function ‘main’.
The function ‘main’ will accept the return value in the variable ‘iSum’ and print the value. We will have
another function called ‘Multiply’ which does not accept any arguments. It just multiplies two values and
prints the result.

#include <stdio.h>

int Sum(int ParamX,

int ParamY)
{
return (ParamX + ParamY);
}

int Multiply(void)
{
int iValue = 25 * 5;
printf(“The multiplication result is %d.\n”, iValue);
}

int main(void)
{
/* Variable Declaration & Initialization */
int iSum = 0;

iSum = Sum(45, 55);

Multiply();
printf(“The sum of the two numbers is %d.\n\n”, iSum);

return 0;
}

The output of the above program will be:

The multiplication result is 125.
The sum of the two numbers is 99.

11.3 FUNCTION VARIABLES: SCOPE AND LIFETIME

The scope and lifetime of a variable depends on its storage class (refer section 2.9). The scope of a
variable determines the parts of the program a variable is available for use. The variables can also be
categorized depending on the place of their declaration. When a variable is declared within a function, it
is referred as a local variable. This variable is available to be accessed and used only within the function
in which it is declared. The lifetime of such a variable depends on the storage class specified for the
variable. When a variable is declared outside every function, it has a global scope and is available to be
accessed and used by every function within the source file (the ‘.c’ file in which it is declared) of the
program. Such a variable is called a global variable and the lifetime of such a variable is equal to the

184
 USER DEFINED FUNCTIONS

entire program lifetime.

11.3.1 AUTOMATIC VARIABLES

Automatic variables are the ones which are declared within a function and need to be created when a
function is called. They are automatically destroyed when the function exits. All variables declared
within a function and without any specified storage class are automatic variables. A variable may also
be specified as automatic by specifying the ‘auto’ keyword in front of its declaration. The ‘auto’ keyword
makes the storage class of the declared variable as automatic. Automatic variables are also referred as
local or internal variables. Following are the examples of automatic variables:
int main(void)
{
int iValue = 0;
auto int iSum;
char szName[100];
double dblPrice;
……..
……..
}

The scope of automatic variables may further be limited to code blocks in which they are declared. Code
blocks are group of statements defined within a set of braces. Consider the below example:

In the previous diagram, we can see that three variables with the same name ‘iValue’ have been
declared but in three different scopes. The variable in scope-1 has its lifetime for the entire ‘main’
function. The second ‘iValue’ has its lifetime only in scope-2. Similarly, the third ‘iValue’ has its lifetime
only in scope-3. The scope of a variable is determined by the innermost curly brace (parenthesis) pair
enclosing the variable. Also, a scope-local variable has higher precedence. So, if we try to access
‘iValue’ within scope-2, the ‘iValue’ declared in scope-2 will be used rather than the one declared in
scope-1. So, printing the value of the variable will print 15. But, the variable declared within scope-2

185
Quick & Indepth C With Data Structures

loses its existence once the control exits scope-2. So, the condition check ‘if (iValue <= 9)’ will do the
comparison based on the value of the variable declared in scope-1 which has its existence for the entire
‘main’ function. But once the control enters the scope-3, the variable declared within scope-3 starts
getting used. If we had not declared any variable with the same name ‘iValue’ (as declared in scope-1)
within scope-3 then, accessing ‘iValue’ (within scope-3) would have meant using the variable declared
in scope-1.

11.3.2 GLOBAL AND EXTERNAL VARIABLES

Variables that are defined outside every function have a lifetime equal to the entire program. These
variables are alive during the entire program execution. Such variables are called global variables.
Unlike local variables, global variables can be accessed by every function in the program.
int iValue;

void PrintValue(void)
{
iValue++;
printf(“The value within PrintValue is %d.\n”, iValue);
}

int main(void)
{
iValue = 32;
printf(“The value within main is %d.\n”, iValue);
PrintValue();
return 0;
}

In the above program, the variable ‘iValue’ is a global variable and is available to both ‘PrintValue’ and
‘main’ functions. The output of the program will be:
The value within main is 32.
The value within PrintValue is 33.

The values stored in global variables can be accessed and changed by any function within the program.
When a function changes a global variable’s value, subsequent functions will start using the new value.
So if there is uncontrolled usage of a global variable within a program, the correctness of the value
stored within it cannot be guaranteed. This may lead to incorrect results. Due to this behavior, we
should try to avoid using global variables as much as possible. For communication between functions,
we should rely more on function arguments (parameters). We should use global variables only in
scenarios where a value is used by multiple functions and argument passing between them is not viable
or possible. We must make sure that the value of this variable is modified in as few places as possible,
though there are minimal or no risks involved for reading such a variable’s value.

Global variables must be declared before they are used within a program. Within the source code file, if

186
 USER DEFINED FUNCTIONS

a function tries to access a global variable which has been declared below it, the compiler will raise an
error. In other words a global variable is visible only from the point of its declaration to the end of the
source code file. Consider the below example:
void PrintValue(void)
{
iValue++;
printf(“The value within PrintValue is %d.\n”, iValue);
}

int iValue;

int main(void)
{
iValue = 32;
printf(“The value within main is %d.\n”, iValue);
PrintValue();
return 0;
}

We have modified our previous program and moved the declaration of ‘iValue’ from above the
‘PrintValue’ function to below it. Due to this change, our program will stop compiling. This is because
when ‘PrintValue’ function tries to access ‘iValue’, the variable ‘iValue’ is still undefined to the compiler.
When the compiler starts compiling the program from top of the source file, it compiles the ‘PrintValue’
function before coming to the declaration of ‘iValue’. As as result, at that stage, the variable ‘iValue’ is
still unknown to the compiler.

Unlike local variables, global variables are automatically initialized to 0 (zero) by the compiler. Hence,
we can start using the global variables without explicitly initializing them.

Now, let us understand the working of global variables using a small program. In the program we have
two global variables ‘g_dblRate’ corresponding to the rate of an item and ‘g_iQuantity’ corresponding to
its quantity. Given these two values, we will calculate total cost of the item.

#include <stdio.h>

/* Global Variables */
double g_dblRate;
int g_iQuantity;

double CalculateCost(void)
{
return g_dblRate * g_iQuantity;
}

void SetRate(void)
{
g_dblRate = 56.90;
}

187
Quick & Indepth C With Data Structures

void AddQuantity(int iQuantity)

{
g_iQuantity += iQuantity;
}

int main(void)
{
g_iQuantity = 50;
SetRate();
printf("Total cost of the item: %.02lf.\n", CalculateCost());

AddQuantity(20);
printf("Total cost after 20 additional quantities: %.02lf.\n", CalculateCost());

return 0;
}

Before explaining external variables, let us discuss a bit about multi-file programs. Till now we have
assumed that all our functions (including main) reside within a single source code file. But, this rarely
happens in an actual project environment. On most occasions there will be multiple programmers
working on a single project and each programmer will be in-charge of a single source code file. This
helps in parallel development of the project, resulting into faster deliveries. All the source code files are
compiled separately and then linked together to form a single executable program.

In a multi-file environment, we may need to share a variable within our program. A global variable
declared in one source file (say source1.c) may need to be accessed or even modified in another source
file (say source2.c). But, there is one complexity in achieving this. When the source file ‘source2.c’ is
compiled, the compiler does not find a declaration of the variable (as the variable is declared in
source1.c) and thus, raises an error. To help compiler know that such a variable is declared in another
source file, we use the keyword ‘extern’. ‘extern’ is derived from the word ‘external’ implying that the
variable is declared external of the source file in which it is getting referenced. The variable is
mentioned as ‘extern’ in the source file in which it is referenced but not declared. When we mention the
variable as ‘extern’, the compiler understands that there is a variable declared with that name and of the
specified data type in some other source file. This variable will be evaluated for existence during link
time rather than compilation time. Let us understand the working with a simple example. Suppose, we
have two source files ‘source1.c’ and ‘source2.c’. ‘g_iCounter’ is a global variable declared in the file
‘source1.c’. Functions within ‘source2.c’ also need access to this variable and thus, this variable is
declared in ‘source2.c’ using the keyword ‘extern’ making the variable available to all the functions in
the source file ‘source2.c’. If the variable is needed only in a few functions within ‘source2.c’, the
‘extern’ declaration can be specified at the function level, within the functions requiring the use of the
variable. So, we can have two possible ways of using the ‘extern’ specifier within ‘source2.c’– at the
global level and at the function level. Both these possibilities are shown in the below table.
Scenario source1.c source2.c
1 int g_iCounter; extern int g_iCounter;

int main(void) void func2(void)

{ {

188
 USER DEFINED FUNCTIONS

Scenario source1.c source2.c

…….. g_iCounter++;
…….. ……..
return 0; ……..
} }

void func1(void) int func3(void)

{ {
g_iCounter++; g_iCounter++;
…….. ……..
…….. ……..
} return 1;
}
2 int g_iCounter; void func2(void)
{
int main(void) extern int g_iCounter;
{ g_iCounter++;
…….. ……..
…….. ……..
return 0; }
}
int func3(void)
void func1(void) {
{ ……..
g_iCounter++; ……..
…….. return 1;
…….. }
}

11.3.3 STATIC VARIABLES

Though we have discussed earlier (section 11.2.4) about static variables, we will look into them again a
little more deeply. Continuing from where we left off in that para, the value of static variables persist till
the end of the program. A variable is declared as static using the keyword ‘static’:
static int iValue;
static double dblValue;
static char szName[100];

A static variable may be declared locally or globally. When declared locally (inside a function), the scope
of the variable lies within the function and cannot be accessed outside it. But, when the function is
called again, the value of the variable continues from where it was left off in the previous call to the
function. The variable is declared and initialized only during the first call to the function. On all later
calls to the function, the variable continues to exist and is not declared or initialized again. We could be

189
Quick & Indepth C With Data Structures

using a static variable to count the number of calls made to the function (possibly during recursion
which we will discuss a little later). A global static variable is declared globally i.e. external to all the
functions within a source file. That variable is available across all the functions. But, there is a catch. A
global static variable is only available to the functions within the source file (‘.c’ file) in which it is
declared. It is not available to functions of other source files within the program. This is a major
difference between a static global variable and a simple global variable.

We can also make a function as static. In such a case, the function becomes visible only to other
functions within the same source file in which it is defined. That function cannot be accessed by
functions in other source files within the program. We declare a function as static by using the keyword
‘static’ in front of the function declaration as below:
static int Func(void);

Now, let us illustrate the use of a static variable using a small program. In the below program, we will
count the number of times a function gets called and print that count.

#include <stdio.h>

void Func(void)
{
static int s_iCallCount = 0;

s_iCallCount++;
printf("Func called %d times.\n", s_iCallCount);
}

int main(void)
{
int iCounter = 0;

for (; iCounter < 5; iCounter++)

Func();

printf("\n");
return 0;
}

Output:
Func called 1 times.
Func called 2 times.
Func called 3 times.
Func called 4 times.
Func called 5 times.

The function ‘Func’ is called by the ‘main’ function 5 times and on each occassion, the function ‘Func’
increments the static counter ‘s_iCallCount’ by 1. The variable gets declared and initialized during the
first call to the function. On all subsequent calls, it does not get declared or initialized again. The
function ‘Func’ then prints the call count. We have initialized the variable ‘s_iCallCount’ with a value of 0
(zero). One important thing to remember is that static variables (like global variables) are automatically

190
 USER DEFINED FUNCTIONS

initialized to 0 (zero), if they are not explicitly initialized. We could have skipped the initialization of the
variable ‘s_iCallCount’ within our program, in which case it would have been implicitly initialized to 0.

11.3.4 REGISTER VARIABLES

In some processor intensive operations (ex: doing several calculations within a time bound period), we
cannot afford to store and retrieve the values of the variables from main memory to processor registers.
Doing store and retrieve repeatedly may affect performance of the operation. In such cases, we may use
register variables which as the name suggests are stored in the processor registers (provided free
register space is available). Register variables are auto by nature but are stored in registers instead of
main memory. Register memory is much faster than main memory, and since all processing is done
using these registers, storing the values in registers improves the performance. We should be careful
when using register variables as register memory is very limited and varies between computer
architectures. We should limit our usage to the minimum and must avoid using register variables
wherever it is not a necessity. When using register variables, we must remember that availability of
register space is not guaranteed and our variables may automatically be shifted to main memory if
register memory is unavailable. We declare register variables within our functions as below:
register Data-Type Variable-Name;

Although, there is no limitation in C regarding the data type of variables that can be placed in registers,
many C compilers restrict the data type to be of int and char. Register variables can only be declared
within functions or code-blocks and cannot be global or external.

11.4 FUNCTIONS WITH POINTER ARGUMENTS

Functions can accept any type of valid data type as their arguments. They can accept simple variables,
structures, unions, arrays and also pointers. Accepting pointers enables a C program to operate faster
and also use less memory. This can be explained with a simple example. Assume, we have two functions
‘func1’ and ‘func2’, where ‘func1’ calls ‘func2’. The program has a structure called ‘MyStruct’ which is
passed as argument to ‘func2’. So, we may code the two functions in the following two ways:
Method 1 Method 2
void func2(struct MyStruct arg) void func2(const struct MyStruct *pArg)
{ {
…….. ……..
…….. ……..
} }

void func1(void) void func1(void)

{ {
struct MyStruct ms; struct MyStruct ms;
…….. ……..
…….. ……..
func2(ms); func2(&ms);
} }

191
Quick & Indepth C With Data Structures

In the first method we are declaring a variable ‘ms’ of type ‘MyStruct’. We are passing the structure
variable as an argument to the call to ‘func2’. In ‘func2’, we declare another variable ‘arg’ of type
‘MyStruct’ as the formal parameter or argument. Now when we pass the actual argument ‘ms’ in the
function call to ‘func2’, the contents of the variable ‘ms’ are assigned to the formal argument ‘arg’. So,
what we have is an assignment of the values of ‘ms’ to the variable ‘arg’. Remember: These two
variables are completely separate variables and changes to one will not affect the others’ values. Hence
in method-1, we have two separate structure variables and the value of one gets assigned to the other.

In the second method we have a little variation. The function ‘func2’ instead of declaring a structure
variable, declares a pointer ‘pArg’ of type ‘MyStruct’. The function ‘func1’ calls ‘func2’ with the address
of the variable ‘ms’ as the actual argument. So instead of the contents of ‘ms’ getting passed to ‘func2’,
the address of ‘ms’ gets assigned to the pointer ‘pArg’ of the function ‘func2’. This makes ‘pArg’ point to
the same memory location of the variable ‘ms’. In method-2, rather than two separate variables
containing the same values, we have one variable ‘ms’ in function ‘func1’, and a pointer ‘pArg’ in ‘func2’
pointing to this variable’s location. So in case of method-2, we not only save the time required for
assigning an entire structure content to another, but also we save memory space as we do not have
separate structure variables but just a pointer to the same memory location.

It should be noted that we have made the ‘pArg’ pointer in method-2 as constant (using the const
keyword). If we had not done so, any changes made to the structure variable in ‘func2’ using the pointer
‘pArg’ would have affected the value contained in ‘ms’ of function ‘func1’. This is because ‘pArg’ is
pointing to the same memory location as ‘ms’ and any changes done using ‘pArg’ would have affected
the contents of ‘ms’. Making ‘pArg’ const prevents function ‘func2’ from making any changes to the
values of the structure.

Whether we should pass the value or the address of a variable to another function depends on the
requirements of the program and must be judged on a case by case basis. Generally for simple data-
type variables, we pass their values if we do not want or require the called function to make any
changes to the variable’s values. If the called function needs to make changes to the values contained
in the actual arguments, we pass their addresses. For data-types such as structures, we generally pass
their addresses to improve the calling efficiency. If we do not want the called function to be making any
changes to the values, we mark the formal parameters as const.

In our next program, apart from 'main', we will have two more functions. One function will accept two
numbers and another will print the numbers.

The first function 'AcceptNumbers' will accept the address of two integer variables. 'AcceptNumbers' will
declare two integer pointers as its formal parameters. The caller (main) function will pass the address of
two integer variables when calling 'AcceptNumbers'. The parameters 'px' and 'py' (in AcceptNumbers)
are declared as pointers which are assigned the addresses of the two variables 'x' and 'y' respectively,
when the function call is made. We pass these same addresses to 'scanf', and 'scanf' fills these memory
locations with the values accepted from the user. So in effect, the values are stored in the variables 'x'

192
 USER DEFINED FUNCTIONS

and 'y' belonging to the function 'main'.

We now pass the values in 'x' and 'y' to the function 'PrintValues'. 'PrintValues' declares two integer
variables 'argX' and 'argY' which are assigned the values of 'x' and 'y' respectively. Note that, changing
the values of 'argX' and 'argY' within 'PrintValues' will have no effect on the values contained in 'x' and
'y' within the 'main' function. This is because we have only assigned the values of 'x' and 'y' to 'argX'
and 'argY' respectively, and not their addresses. So, 'argX' and 'argY' only get a copy of the values
contained in 'x' and 'y'.

After printing the two numbers, 'PrintNumbers' adds the numbers and returns their sum as an integer
(int). The 'main' function accepts the sum of the two numbers (returned by 'PrintNumbers') in the
variable 'iSum' and prints its value.

#include <stdio.h>

void AcceptNumbers(int *px,

int *py)
{
/* Check if the caller function has sent us valid addresses. */
if (px == NULL || py == NULL)
{
return;
}

printf("Enter the first value: ");

scanf("%d", px);

printf("Enter the second value: ");

scanf("%d", py);
}

int PrintNumbers(int argX,

int argY)
{
printf("\nThe two numbers are %d and %d.", argX, argY);
return (argX + argY);
}

int main(void)
{
/* Variable Declaration & Initialization */
int x = 0;
int y = 0;
int iSum = 0;

AcceptNumbers(&x, &y);
iSum = PrintNumbers(x, y);
printf("\nThe sum of the two numbers is %d.\n\n", iSum);

return 0;
}

11.4.1 FUNCTION POINTER AS ARGUMENT

Just like any variable, a function also has a specific address of its own. This is the starting address of
the location where the function’s binary code is loaded in memory for execution. Similar to a variable,

193
Quick & Indepth C With Data Structures

we can pass the address of a function to another function. Within the called function we declare a
pointer as the formal argument, which accepts the address of the function. This helps in scenarios
where we want the called function to call another function (say func3) for further processing. The other
function that needs to be called varies depending on the logic. We will explain this with a simple
program.
#include <stdio.h>

/* Define a function type pointer and give it a name. */

typedef double (*PFNCALCROUTINE)(double, double);

/* Function for adding to numbers. */

double sum(double x, double y)
{
return x + y;
}

/* Function to find the difference between two numbers. */

double subtract(double x, double y)
{
return x – y;
}

/* Function to find the product of two numbers. */

double multiply(double x, double y)
{
return x * y;
}

/* Function to divide one number by another. */

double divide(double x, double y)
{
return x / y;
}

/* The intermediary function which calls any function which accepts two arguments
of type double and returns a double value. It calls the function pointed by
the function pointer 'pfn' with the arguments 'x' and 'y'. 'Calculate' returns
the same value returned by the function pointed by 'pfn'. */
double Calculate(double x,
double y,
PFNCALCROUTINE pfn)
{
if (pfn)
return pfn(x, y);
return 0;
}

int main(void)
{
double dblNum1 = 0;
double dblNum2 = 0;

printf("Enter the first number: ");

scanf("%lf", &dblNum1);

printf("Enter the second number: ");

scanf("%lf", &dblNum2);

/* Call the function 'Calculate' with the appropriate function pointer. 'Calculate'

194
 USER DEFINED FUNCTIONS

will call the function using the provided function pointer. */

printf("\nThe sum of the numbers is %lf.", Calculate(dblNum1, dblNum2, &sum));
printf("\nThe difference of the numbers is %lf.", Calculate(dblNum1, dblNum2, &subtract));
printf("\nThe product of the numbers is %lf.", Calculate(dblNum1, dblNum2, &multiply));
printf("\nThe division of the numbers is %lf.\n\n", Calculate(dblNum1, dblNum2, &divide));

return 0;
}

In our above program we are telling the compiler that we will have functions which will accept two
arguments of type ‘double’. Such functions will also return a value of type ‘double’. We are type defining
pointers of such function types as ‘PFNCALCROUTINE’. We are defining four worker functions ‘sum’,
‘subtract’, ‘multiply’ and ‘divide’ which perform the actual operations (calculations in this case). The
function ‘Calculate’ is an intermediary function which is responsible for calling these worker functions
and getting the job done by them. Whatever is returned by the worker function is in turn returned by
‘Calculate’. Within the function ‘main’, we are actually calling the function ‘Calculate’ and passing it the
address of the worker function which is responsible for performing the actual calculation. ‘Calculate’ in
turn calls the function pointed by the function pointer ‘pfn’.

11.5 FUNCTIONS WITH ARRAY ARGUMENTS

Till now we have discussed passing single valued arguments to functions. If we have array arguments to
pass to another function, we may pass the address of the starting location of the arrays. The called
function then works directly on the memory location of the array. If we do not want the called function
to be making changes to the array, we mark the formal parameter of the called function as ‘const’.

The called function works using the pointer to the starting location of the array, but, how does it know
the size of the array? If it does not know the size of the array, it may easily overshoot the array index or
try to access an element which does not exist. We can prevent this using two mechanisms. In the first
mechanism, the called function may always accept a fixed size array by declaring a pointer to the entire
array (refer section ‘Pointer to an array’ of section 8.4). This will prevent any caller function from
passing an array of a different size other than the one specified as the formal argument of the called
function. In the second mechanism, we could pass the address to the first element of the array i.e. the
starting location of the array. We also pass another argument with the array size (number of elements in
the array). This will make the called function work according to the count of elements in the array.
Consider the below example explaining the two mechanisms.

#include <stdio.h>

#define MAX_ELEMENTS 10

void AcceptNumbers(int *prgValues,

int iNumElements)
{
int iCounter = 0;

if (!prgValues)
{
return;
}

195
Quick & Indepth C With Data Structures

for (; iCounter < iNumElements; iCounter++)

{
printf("Enter the number: ");
scanf("%d", &prgValues[iCounter]);
}
}

void PrintNumbers(const int *prgValues,

int iNumElements)
{
int iCounter = 0;

if (!prgValues)
{
return;
}

printf("\nThe numbers are: ");

for (; iCounter < iNumElements; iCounter++)
{
printf("%d", prgValues[iCounter]);
if ((iCounter + 1) != iNumElements)
printf(", "); /* We do not want a comma after the last number. */
}
}

double FindAverage(const int (*prgValues)[MAX_ELEMENTS])

{
int iCounter = 0;
int iSum = 0;

if (!prgValues)
{
return 0;
}

for (; iCounter < MAX_ELEMENTS; iCounter++)

{
iSum += (*prgValues)[iCounter];
}

return (double)iSum / MAX_ELEMENTS;

}

int main(void)
{
int irgValues[MAX_ELEMENTS] = { 0 };

AcceptNumbers(irgValues, MAX_ELEMENTS);
PrintNumbers(irgValues, MAX_ELEMENTS);
printf("\n\nThe average of the numbers is %.02lf.\n\n", FindAverage(&irgValues));

return 0;
}

In the pointer ‘prgValues’, the function ‘AcceptNumbers’ accepts the address of the starting location of
the array. It also accepts the element count of the array. The function fills the elements of the array with
the values accepted from the user. The number of values accepted depends on the element count of the
array which it received from the caller in ‘iNumElements’. If we are building functions which are flexible

196
 USER DEFINED FUNCTIONS

and are not targeted towards arrays with some specific number of elements, we need the functions to
behave like this. The arguments of function ‘PrintNumbers’ are very similar to ‘AcceptNumbers’ except
that we are declaring the pointer to be of type ‘const’. This prevents the function from making any
changes to the element values.

The function ‘FindAverage’ is a little different in its prototype or declaration. Rather than declaring a
pointer to the first element of the array as argument, it declares a pointer to the entire array. This
enables it to specify the number of elements (MAX_ELEMENTS in our case) it expects within the array.
So, any caller function trying to specify an array of a different size (other than MAX_ELEMENTS) will
generate a compilation error. This makes the function targeted specifically for arrays with
MAX_ELEMENTS number of elements and saves ourselves from specifying the element count as a
separate parameter to the function.

11.6 FORWARD DECLARATION OF FUNCTIONS

When a source file is compiled by the compiler, every function prototype (signature) i.e. its name,
arguments and return type is added to its function dictionary. The compiler compiles the program from
top to bottom and proceeds downwards by adding each function signature to the dictionary. When it
encounters a function call, it searches its dictionary for the function signature and if it finds one, it
continues compiling. If it does not find the required function signature, it raises an error and may stop
the compilation. So if in our source file, we have a function call specified before the function’s definition,
the compiler (mainly single-pass compilers) will fail to find the function signature in its dictionary and
will raise an error. We may encounter situations where it becomes necessary to define a function a little
later. For such situations, we may take the help of forward declarations. In forward declarations, we
specify the function prototype or signature of a function before the first call to the function is made
within the source file. The definition or body of the function may come later within the source file. When
this is done, the compiler understands that the function is defined somewhere else within the program
and adds the signature to its dictionary. The check for the function definition is postponed to link time.
Hence, calls to the function before its definition will not raise an error. The function prototype is
specified as below:
Return-Data-Type Function_Name(List of arguments);

Let us understand this with a very simple example:

#include <stdio.h>

/* Forward Declaration. The call to the 'StandardDeviation' function

is made before the function is defined. When the compiler encounters
the call (in the function 'main'), it still does not have the function
signature in its function list. This lets the compiler know what the
function accepts as arguments and what it is supposed to return. */
double StandardDeviation(const int *, int);

int main(void)
{
int irgValues[10] = { 78, 26, 8, 17, 190, 3, 710, 62, 10, 82 };

printf("Standard Deviation of the numbers is %lf.\n\n",

StandardDeviation(irgValues, 10));

197
Quick & Indepth C With Data Structures

return 0;
}

double StandardDeviation(const int *prgValues,

int iNumElements)
{
int iCounter = 0;
double dblSum = 0;
double dblMean = 0;
double dblDiff = 0;
double dblVariance = 0;

if (!prgValues || iNumElements <= 1)

{
return 0;
}

/* Find the sum of all the numbers. */

for (; iCounter < iNumElements; iCounter++)
{
dblSum += prgValues[iCounter];
}

/* Now, find the mean. */

dblMean = dblSum / iNumElements;

/* Find the sum of the squared deviation from the mean. */

for (iCounter = 0; iCounter < iNumElements; iCounter++)
{
dblDiff = prgValues[iCounter] – dblMean;
dblVariance = dblVariance + dblDiff * dblDiff;
}

/* The variance... */
dblVariance = dblVariance / (iNumElements – 1);

/* Return the standard deviation. */

return sqrt(dblVariance);
}

In the above program, we are finding the sample standard deviation of an array of 10 numbers. We have
defined the ‘StandardDeviation’ function after the function ‘main’ where the call to the
‘StandardDeviation’ function is made. So, we have specified the signature of the function
‘StandardDeviation’ prior to the function ‘main’ to let the compiler add the function signature to its
dictionary before the call to the function is made within the ‘main’ function. The sample standard
deviation of a group of numbers (x1, x2, x3, ……., xn) is given as:

where, is the mean of all the numbers. We first find the mean of all the numbers. Next, we find the
square of the difference of each number from the mean. We also find the summation of all these squares
and divide the summation by ‘n – 1’. What we now have is the variance, and the square root of the
variance is the standard deviation.

198
 USER DEFINED FUNCTIONS

11.7 VARIABLE NUMBER OF ARGUMENTS

We have extensively used the C standard function ‘printf’, which accepts variable number of arguments.
If we want a similar function of our own which accepts an indeterminate number of arguments, how do
we achieve it? Assume we define a function which accepts any number of integer values and returns the
average of the values. We can achieve this using a function which accepts variable number of
arguments. Variable number of arguments is denoted by the use of ellipse (given as ‘...’). The ellipse
must be the last argument of the function and the function must have at-least one argument preceding
the ellipse.

We can read the arguments using a set of macros (‘va_start’, ‘va_arg’ and ‘va_end’). A macro is a
fragment of code specified using the ‘#define’ statement, which has been given a name. Wherever the
name is used, it is replaced by the contents of the macro following the defined name. Consider the
below macro:
#define ARRAY_SIZE(x) sizeof(x)/sizeof(x[0])

The above macro calculates the number of elements in an array. We call the macro with an array name
as below:
int rgValues[10] = { 0 };
int iNumElements = ARRAY_SIZE(rgValues);

The macro divides the total size of the array by the size of the first element of the array returning the
number of elements in the array. Wherever the macro name is specified, it is replaced by the code
following the macro name (i.e. ‘sizeof(x)/sizeof(x[0])’), where ‘x’ is replaced by the value or variable
specified in the macro call (‘rgValues’ in our case).

The three macros ‘va_start’, ‘va_arg’ and ‘va_end’ have the following signatures:
void va_start(va_list arg_ptr, prev_param);
type va_arg(va_list arg_ptr, type);
void va_end(va_list arg_ptr);

where ‘va_list’ is a character pointer which is made to point to the next argument in the variable list.
‘va_start’ initializes the ‘arg_ptr’ to point to the first argument in the variable number of argument list.
‘va_arg’ makes ‘arg_ptr’ point to the next argument in the variable number of argument list. It also
returns the current argument pointed by ‘arg_ptr’ after formatting it to the data type specified as the
second argument to the macro. ‘va_end’ must be specified after we have completed processing our
variable arguments list.

The prototype or signature of a function which accepts variable number of arguments is as below:
Return-Data-Type function_name(Fixed Arguments, …);

Now, let us understand the use of variable number of arguments using a small program. We will write a
function which accepts an indeterminate number of integer values. These values are specified after an

199
Quick & Indepth C With Data Structures

argument ‘iArgCount’. ‘iArgCount’ specifies the count of the variable number of arguments specified in
the call to the function. The function finds the average of all the specified values.
#include <stdio.h>
#include <stdarg.h>

double FindAverage(int iArgCount, …)

{
va_list valist = NULL;
double dblSum = 0;
int idx = 0;

/* Initializes 'valist' to have the first optional argument specified

after 'iArgCount'. As the second argument, we pass the preceding
argument to the ellipse to our function. */
va_start(valist, iArgCount);

/* Access all the variable number of arguments one at a time. */

for (idx = 0; idx < iArgCount; idx++)
{
/* After the call, 'valist' is set to the next argument in the
list. 'va_arg' returns the current content in 'valist' formatted
to the data-type specified in the second argument. */
dblSum += va_arg(valist, int);
}

/* Clean memory reserved for 'valist'. */

va_end(valist);

return dblSum / iArgCount;

}

int main(void)
{
/* The first argument ‘5’ to ‘FindAverage’ is the count of the variable
number of arguments following this argument. */
printf("The average of the numbers is %lf.\n\n",
FindAverage(5, 2, 4, 6, 8, 10));

return 0;
}

We are including the header file ‘stdarg.h’ as the macros ‘va_start’, ‘va_arg’ and ‘va_end’ are defined in
this header file.

11.8 RECURSION
Think about a group of statements that needs repeated execution. The first thing that comes to our
mind is to use looping constructs. What if, the group of statements almost spans the entire function. For
such situations, we may use recursion where a function calls itself.

When a function calls itself, the new instance of the function executes in a manner as if a fresh new call
is made to the function. This means that all the non-static variables of the function start afresh for the
new call and are again re-declared and re-initialized for that call instance. The previous call instance
(which made the new call) continues carrying its own set of variable values. In simple terms, each called
instance of the function carries its own set of variable values distinct from the other instances. Each

200
 USER DEFINED FUNCTIONS

called instance starts executing from the first statement of the function. A simple example of recursion
is given here under:
#include <stdio.h>

#define TRIANGLE_HEIGHT 10

void RightTriangle(int iMaxStars)

{
/* For every called instance of the function, we have a separate value of iCounter */
int iCounter = 0;

/* Print the stars up to a count of iMaxStars.

In this function, we are printing our line first before asking the next
call to the function to print the next line. So, 1 star is printed first,
followed by the line with 2 stars, then the line with the 3 stars and so
on. The line with 10 stars is printed last, giving us an impression of a
right-angled triangle. */
for (; iCounter < iMaxStars; iCounter++)
{
printf("*");
}

/* Make the next set of stars come in the next line. */

printf("\n");

/* After printing the 10th line of stars, we stop calling the function again.
This is the exit criteria which makes the recursion stop and the function
instances start exiting with the last function instance stopping/exiting
first followed by the previous-last and so on. */
if (iMaxStars < TRIANGLE_HEIGHT)
RightTriangle(iMaxStars + 1);
}

void InvertedRightTriangle(int iMaxStars)

{
int iCounter = 0;

/* Before printing our own line, we ask the next function instance to draw
its line. That instance again does the same. This continues until we are
at the instance which needs to draw the longest line with 10 stars. We
do not call further and print our line with 10 stars. We then print a
newline and then exit this instance. For the previous instance, the
recursive call to 'InvertedRightTriangle' ends and it prints its own line
with 9 stars. It too prints a newline and exits. This continues till the
first instance has printed its line with 1 star and exits. So, in this
mechanism, the last instance prints its line first, followed by the
previous-last and so on. The first instance prints its line the last. */
if (iMaxStars < TRIANGLE_HEIGHT)
InvertedRightTriangle(iMaxStars + 1);

for (; iCounter < iMaxStars; iCounter++)

{
printf("*");
}

printf("\n");
}

int main(void)
{
RightTriangle(1);

201
Quick & Indepth C With Data Structures

InvertedRightTriangle(1);

return 0;
}

The output of the above program will be:

*
**
***
****
*****
******
*******
********
*********
**********
**********
*********
********
*******
******
*****
****
***
**
*

It is important to remember that recursion operates in LIFO mechanism i.e. Last In First Out. This means
that the function instance which was invoked the last, is the first one to complete its execution (end).
The first instance to be invoked is the last one to end.

Recursion is used to solve problems where we need to apply the same solution successively to subsets
of the same problem. This means the bigger problem is broken up into multiple subsets with each
subset being a similar representation of the bigger problem. When writing a recursive function we need
to be careful in placing a proper exit criteria which effectively ends the recursion. If we miss the exit
criteria, the recursion will continue indefinitely until we run out of memory. We should also be careful
regarding the positioning and correctness of the exit criteria, without which we will generate incorrect
results. Exit criteria determines the moment when we stop calling ourselves further, which effectively
starts the end of the recursion. If the exit criteria is not met, the recursion continues. The recursion
starts wrapping up when the exit criteria is hit. In our above program the condition check ‘if (iMaxStars
< TRIANGLE_HEIGHT)’ acts as the exit criteria. If the condition is fulfilled we continue the recursion, and
we stop when the condition is not met. We may also call it the recursion criteria, as we continue the
recursion depending on the condition being met.

Look carefully at the positioning of the recursion criteria in our above program. The recursion criteria is
placed differently in the functions ‘RightTriangle’ and ‘InvertedRightTriangle’. This makes the two
functions behave differently, where the former generates an upside right-angled triangle and the latter
generates a downward right-angled triangle. The function ‘RightTriangle’ draws its own line first and

202
 USER DEFINED FUNCTIONS

then calls itself recursively. This makes the line with 1-star being drawn first, followed by line with 2-
stars and so on. The function ‘InvertedRightTriangle’ calls itself recursively before drawing its own line.
This makes the last called instance to draw its line with 10-stars first, followed by the call which draws
the line with 9-stars and so on.

Recursion can be of two types: Direct recursion and indirect recursion. Our above program is an example
of direct recursion where the function directly calls itself. In case of indirect recursion, the function calls
another function which in turn calls this function. As for example, function f1 calls function f2 which in
turn calls the function f1, and this continues recursively. Indirect recursion is also called mutual
recursion, as two or more functions are calling each other creating a recursive behavior.

11.9 SAMPLE PROGRAMS

1. In our next program we will print a geometric progression. We will accept the starting number
for the sequence and the common ratio. We will print a maximum of 10 numbers in the
sequence.

We are using recursion where the function 'PrintGP' is calling itself to get the next number in
sequence printed. The recursion continues until we have printed all the 10 numbers. The exit
condition of the recursion is whether we have reached the 10 numbers, upon which we start
coming out of the recursion. We have declared a global variable 'g_iNumCount' which keeps a
count of the numbers we have printed.

/* Header Files */
#include <stdio.h>

/* Defined Values */
#define MAX_NUMBERS 10

/* Global variables */
int g_iNumCount;

void PrintGP(long int iCurNum, int iCommonRatio)

{
/* Variable Declaration & Initialization */
long int iNextNum = 0;

/* Print the current number in sequence. */

printf("%d ", iCurNum);
g_iNumCount++;

/* Have we already printed all the numbers. If so, do not

print any more and start coming out of the recursion. */
if (g_iNumCount < MAX_NUMBERS)
{
/* Find the next number in sequence and ask the next
recursive call to 'PrintGP' to print that number. */
iNextNum = iCurNum * iCommonRatio;
PrintGP(iNextNum, iCommonRatio);
}
}

203
Quick & Indepth C With Data Structures

int main(void)
{
/* Variable Declaration & Initialization */
int iStartNum = 0;
int iCommonRatio = 0;

printf("Enter the starting number for the geometric sequence: ");

scanf("%d", &iStartNum);

GETCOMMONRATIO:
/* Refrain from giving a large number as we may overflow.
To save ourselves from overflowing, we may have to provide
checks on the value of common ratio and also use Int64. */
printf("Enter the common ratio: ");
scanf("%d", &iCommonRatio);

if (iCommonRatio <= 0)
{
goto GETCOMMONRATIO;
}

g_iNumCount = 0;

printf("\n\nThe Geometric sequence is:\n");

PrintGP(iStartNum, iCommonRatio);
printf("\n\n");

return 0;
}

2. In our next program we will touch on the subject of compound interest for bank deposits.
Compound interest is that interest which is added back to the principal sum so that the interest
is earned on that added interest during the next compounding period. The shorter the
compounding period the higher will be the maturity amount.

Compounding frequency is the frequency at which the interest is calculated (ex. every month,
every quarter, every twelve months). A monthly compounding (compound frequency of 12) will
earn a higher maturity amount than a half-yearly compounding (compound frequency of 2), as
the interest on interest is calculated much more frequently.

The formula for calculating maturity amount based on compound interest is:
(n * t)
A = P * (1 + r / n)
Where:
A → The amount receivable including interest after the maturity period.
P → The principal investment amount (the initial deposit).
r → The annual rate of interest.
n → The number of times the interest is compounded per year.
t → The period of interest.

The amount of interest that will be earned is calculated as 'I = A - P', where 'I' is the interest
earned.

204
 USER DEFINED FUNCTIONS

In the program we will accept the principal investment amount, annual rate of interest,
compounding frequency and the period of investment. We will then calculate the total maturity
amount and the total interest earned.

#include <stdio.h>

/* Structure Definition */
typedef struct _CMPDINT
{
double dblPrincipal;
float fltInterestRate;
int iCompoundFrequency;
float fltPeriod;
double dblMaturityAmount;

}CMPDINT;

CMPDINT GetValues(void)
{
/* Variable Declaration & Initialization */
CMPDINT cmpdint = { 0 };

printf("Enter the principal investment amount: ");

scanf("%lf", &cmpdint.dblPrincipal);

printf("Enter the annual rate of interest: ");

scanf("%f", &cmpdint.fltInterestRate);

while (cmpdint.iCompoundFrequency > 12 || cmpdint.iCompoundFrequency < 1)

{
printf("Enter the compound frequency (1 - 12): ");
scanf("%d", &cmpdint.iCompoundFrequency);
}

printf("Enter the period of interest (in years): ");

scanf("%f", &cmpdint.fltPeriod);

return cmpdint;
}

void CalculateMaturity(CMPDINT *pcmpdint)

{
/* Variable Declaration & Initialization */
double dblRate = 0;
double dblPeriod = 0;

if (pcmpdint)
{
/* fltInterestRate is in percentage. Convert to decimals. */
dblRate = (double) pcmpdint->fltInterestRate / 100;

dblRate = 1 + (dblRate / pcmpdint->iCompoundFrequency);

dblPeriod = pcmpdint->iCompoundFrequency * (double) pcmpdint->fltPeriod;

pcmpdint->dblMaturityAmount = pcmpdint->dblPrincipal * pow(dblRate, dblPeriod);

}
}

205
Quick & Indepth C With Data Structures

int main(void)
{
/* Variable Declaration & Initialization */
CMPDINT cmpdint = { 0 };

cmpdint = GetValues();
CalculateMaturity(&cmpdint);

printf("\nThe maturity value is %lf and the total interest if %lf.\n\n",

cmpdint.dblMaturityAmount, cmpdint.dblMaturityAmount – cmpdint.dblPrincipal);

return 0;
}

3. In our next program we will try to find whether a given number is a Narcissistic number. A
Narcissistic number is a number that is equal to the sum of each of its digits raised to the power
of the number of digits in the number. For example: '153 = 13 + 53 + 33, is a Narcissistic number.

/* Header Files */
#include <stdio.h>

/* Function which counts the number of digits in a number. */

int CountDigits(int iNumber)
{
/* Variable Declaration & Initialization */
int iNumDigits = 0;

while (iNumber > 0)

{
iNumber = iNumber / 10;
iNumDigits++;
}

return iNumDigits;
}

/* Function which determines if a number is Narcissistic or not. */

int IsNarcissistic(int iNumber)
{
/* Variable Declaration & Initialization */
int iNumDigits = 0;
int iModulus = 0;
int iSum = 0;
int iTemp = iNumber;
int fNarcissistic = 0;

/* Count the number of digits in the number. */

iNumDigits = CountDigits(iNumber);

/* Calculate through all the digits of the number. */

while (iTemp > 0)
{
iModulus = iTemp % 10;
iTemp = iTemp / 10;

iSum = iSum + (int) pow(iModulus, iNumDigits);

}

206
 USER DEFINED FUNCTIONS

if (iSum == iNumber)
{
fNarcissistic = 1;
}

return fNarcissistic;
}

int main(void)
{
/* Variable Declaration & Initialization */
int iNumber = 0;

printf("Enter the number to check: ");

scanf("%d", &iNumber);

if (0 != IsNarcissistic(iNumber))
{
printf("\nThe number %d is a Narcissistic number.\n\n",
iNumber);
}
else
{
printf("\nThe number %d is not a Narcissistic number.\n\n",
iNumber);
}

return 0;
}

4. The Tower of Hanoi (also called the Tower of Brahma or Lucas Tower) is a mathematical puzzle.
It consists of three pegs and a number of disks of different sizes, which can slide onto any of
the pegs. The puzzle starts with the disks in ascending order of size stacked on one peg with
the smallest at the top, thus creating a conical shape.

The objective of the puzzle is to move the entire stack to another peg, obeying the following
rules:
◦ Only one disk can be moved at a time.
◦ In each move we can take the topmost disk from one of the stacks and place it on top of
another stack.
◦ No disk may be placed on top of a smaller disk.

Assume we label the pegs as 'PegA', 'PegB' and 'PegC', and we have 'n' disks stacked on 'PegA'
with '1' being the smallest and 'n' being the largest disk. We want to move the disks from 'PegA'
to 'PegC' by using 'PegB' as an intermediary/auxiliary peg.

We break down the problem into smaller sub-problems and solve the sub-problems individually.
This makes the larger problem getting solved automatically. If we work with just 3 disks, what
will we do? We will do the following 7 steps:
1. Transfer the smallest disk '1' from 'PegA' to 'PegC'.
2. Transfer the disk '2' from 'PegA' to 'PegB'.

207
Quick & Indepth C With Data Structures

3. Transfer the disk '1' from 'PegC' to 'PegB'.

4. Transfer the disk '3' from 'PegA' to 'PegC'.
5. Transfer the disk '1' from 'PegB' to 'PegA'.
6. Transfer the disk '2' from 'PegB' to 'PegC'.
7. Transfer the disk '1' from 'PegA' to 'PegC'.

For a total of 'n' disks, there will be a minimum of (2 n - 1) transfers required to achieve the
result.

For every disk 'm', the general mechanism for solving the problem is:
1. Move (m − 1) disks from source to auxiliary, by the same general solving procedure.
2. Move the disk 'm' from the source to the target peg.
3. Move the (m − 1) disks that we have just placed on the auxiliary, from the auxiliary to the
target peg by the same general solving procedure.

When working with mth disk, the above three steps are repeated recursively for each of the disks
till ‘m – 1’, starting from the smallest disk to the largest.

Now let us implement a Tower of Hanoi using three structure variables. Each structure variable
will be a logical or computer representation of a single peg in the Tower of Hanoi. The structure
on which these variables will be based will have three members:
1. The first member will be a string which will identify the structure variable to the peg in the
Tower of Hanoi. This will be done using a name assigned to it (ex: PegA, PegB).
2. The second member will be an array of integers which will hold the values aka. disks
stacked on that peg. The larger the number the bigger the disk size. So as per the rules, the
numbers will appear in descending order within the array.
3. The third member will hold the number of values aka. disks currently stored on the peg
represented by this structure variable.

Notice that during the first recursive call, the current auxiliary peg becomes the target. In the
second recursive call, the current auxiliary becomes the source.

#include <stdio.h>

/* Defined Values */
#define MAX_DISKS 4

/* Structure Definition and global variables */

struct HanoiPeg
{
char szName[20];
int rgiDisks[MAX_DISKS];
int iCurNum;
}
g_Src = { "PegA", { 0 }, 0 },
g_Aux = { "PegB", { 0 }, 0 },
g_Target = { "PegC", { 0 }, 0 };

208
 USER DEFINED FUNCTIONS

void TowerOfHanoi(int n,
struct HanoiPeg *pSrc,
struct HanoiPeg *pTarget,
struct HanoiPeg *pAux)
{
/* Variable Declaration & Initialization */
int idx = 0;

if (n > 0)
{
/* Move all smaller disks from source to auxillary
and then from auxillary to target. */
TowerOfHanoi(n - 1, pSrc, pAux, pTarget);

pSrc->rgiDisks[--pSrc->iCurNum] = 0;
pTarget->rgiDisks[pTarget->iCurNum++] = n;

/* Print our progress. */

printf("\n################################################");

printf("\nMoving disk %d from %s to %s", n, pSrc->szName, pTarget->szName);

/* Over here the source may not be PegA, auxillary may not be PegB and
the target may not be PegC. But, when printing we need to keep the
order PegA, then PegB and then PegC for better readability. */
printf("\nPeg %s: ", g_Src.szName);
for (idx = 0; idx < g_Src.iCurNum; idx++)
printf("%d ", g_Src.rgiDisks[idx]);

printf("\nPeg %s: ", g_Aux.szName);

for (idx = 0; idx < g_Aux.iCurNum; idx++)
printf("%d ", g_Aux.rgiDisks[idx]);

printf("\nPeg %s: ", g_Target.szName);

for (idx = 0; idx < g_Target.iCurNum; idx++)
printf("%d ", g_Target.rgiDisks[idx]);

printf("\n################################################\n");

/* Move all smaller disks from source to auxillary

and then from auxillary to target. */
TowerOfHanoi(n - 1, pAux, pTarget, pSrc);
}
}

int main(void)
{
/* Variable Declaration & Initialization */
int idx = 0;
int IDiskNum = MAX_DISKS;

for (; idx < MAX_DISKS; idx++, IdiskNum--)

{
g_Src.rgiDisks[idx] = IdiskNum;
}

g_Src.iCurNum = MAX_DISKS;
TowerOfHanoi(MAX_DISKS, &g_Src, &g_Target, &g_Aux);

printf("\n");
return 0;
}

209
Quick & Indepth C With Data Structures

FILE MANAGEMENT

[CHAPTER-12]

12.1 INTRODUCTION
Till now we have been working with console oriented input and output of data to a program. Our
programs accept input from the user, processes them and shows the output on the screen. This works
for data which is limited in size and does not require to be referenced later. What happens if we require
the processed data to be available at a later time. Do we input the raw data and run the program to
process the data again? If we had the option to store the processed data, we would not have required to
enter the raw data again and could have referenced the stored data during our later use. This is where
storing the data on disks comes in handy. The processed data is stored in files which are saved to disks.
On a later use, the program just reads the data from the file and shows it to the user. This fulfills the
following two purposes:
1. We do not need to input the entire data again to the program. Saving in files helps eliminate this
cumbersome operation.
2. The program does not need to do the entire processing again. It just needs to read the already
processed data from the files and show to the user.

The various file operations supported by C are:

• Creating a new file.
• Opening an existing file.
• Writing to a file.
• Reading a file.
• Modifying a file.
• Overwriting or clearing a file.
• Closing a file.

We can open and operate on a file in two different formats – Textual and Binary. In textual format the
text and characters are stored one character a time. But, numbers are stored as strings of characters.
Even if the number 637283 occupies 4 bytes of memory, when transferred to the disk in textual format,
would occupy 6 bytes, with each digit as a separate character. So, each data item is stored as an ASCII
or Unicode character (depending on the C function used). In case of binary format, the exact memory
representation of the data is written to disk. So, 637283 will use 4 bytes when using binary format.

We may use many of the C library functions to perform the file operations, and they are:
Function Name Function Description
fopen Creates a new file or opens an existing file.
fclose Closes a file.
fseek Sets the position to a desired byte location within the file. The next read
or write operation begins from this location.

210
 FILE MANAGEMENT

Function Name Function Description

ftell Returns the current position within the file (in terms of bytes from the
start of the file).
rewind Sets the position to the beginning of the file. The next read or write
operation begins from the start of the file.
getw Reads an integer value from a file.
putw Writes an integer value to a file.
ANSI → getc Reads a character from a file.
Wide Character → fgetwc
ANSI → putc Writes a character to a file.
Wide Character → fputwc
ANSI → fgets Reads a character string from a file. Reads until a new line character or
Wide Character → fgetws end of file is reached (whichever comes first).
ANSI → fputs Writes a character string to a file. Writes until the NULL character, but
Wide Character → fputws does not write the NULL character itself to the file.
ANSI → fprintf Writes a set of formatted data values to a file. Similar to ‘printf’, except
Wide Character → fwprintf that it can write to the specified output stream which may be a file.
ANSI → fscanf Reads a set of formatted data values from a file. Similar to ‘scanf’, except
Wide Character → fwscanf that it can read from the specified input stream which may be a file.

12.2 CREATING AND OPENING FILE

A file in secondary storage is identified by its name and is accessed using its file path. For creating or
opening a file, we must specify the path to the file. The path may be the full path to the file or simply the
filename. When only the filename is specified, the file is assumed to be in the current working directory.
For creating or opening a file, we may use the ‘C’ function ‘fopen’. To the function, we must specify a
filename or file path which satisfies the operating system’s file naming and path conventions. The
function returns a pointer to a data structure called ‘FILE’. The function has the following prototype:
FILE * fopen(“file path”, “mode”);

As the name suggests, ‘file path’ is a character string representing the path to the file to create or open.
It may also be a filename in which case the file is assumed to be in the current working directory. The
parameter ‘mode’ suggests the purpose of opening the file and may be the following:
r Opens the file for reading only. If the file does not exist, the function fails and returns NULL.
After opening, the file read-position is stationed at the beginning of the file.
w Opens the file for writing only. If the file does not exist, a new file is created. If the file already
exists, it is overwritten and no data exists in the file after opening. After opening, the file
write-position is stationed at the beginning of the file.
a Opens the file only for appending (adding) data. If the file does not exist, a new file is created.
If the file already exists, it is opened (not overwritten) and the file write-location is stationed
at the end of the file for appending data to it.
r+ Opens the file for both reading and writing. If the file does not exist, a new file is created. If
the file already exists, it is opened and the file write-location is stationed at the start of the
file. Previous data may be overwritten, if needed.

211
Quick & Indepth C With Data Structures

w+ Same as ‘w’ except that the file is opened for both reading and writing.
a+ Same as ‘a’ except that the file is opened for both reading and writing.

Additionally, we have the option to specify the format (i.e. Textual or Binary) when opening the file. The
default format is textual. If we want to open a file in binary format, we have to specify an additional flag
‘b’ in the mode parameter of ‘fopen’. If ‘b’ is not specified, the file is opened for textual operation.
Starting from C2011, another flag ‘x’ may be specified in the mode parameter. This flag forces the
function to create a new file rather than overwrite any existing file. If a file with the same name already
exists, the function ‘fopen’ fails, rather than opening the file. If we use ‘x’, we can be sure that we will
not clobber an existing file. Some usage examples of ‘fopen’ are given below:
Variable Declaration: FILE *fp = NULL;
Function Usage Examples: 1. fp = fopen(“c:\\data\\myfile.dat”, “r”);
2. fp = fopen(“c:\\data\\myfile.dat”, “a”);
3. fp = fopen(“c:\\data\\myfile.dat”, “r+”);
4. fp = fopen(“c:\\data\\myfile.dat”, “w+”);
5. fp = fopen(“c:\\data\\myfile.dat”, “ab”);
6. fp = fopen(“c:\\data\\myfile.dat”, “r+b”);
7. fp = fopen(“myfile.dat”, “wb”);

The function ‘fopen’ returns a pointer to a ‘FILE’ data structure. This data structure holds all
information regarding the opened file. The returned pointer is subsequently used for all operations on
the opened file.

We can open and use multiple files at the same time. The maximum number of files that can be opened
depends on the operating system restrictions.

By default, text files opened using ‘fopen’ store data in ANSI format. If we want the files to store the
characters in some other encoding like ‘UTF-8’ or ‘UTF-16’, we need to specify some additional
attributes within the ‘mode’ parameter. If the 'mode' string contains the additional sequence
',ccs=encoding', 'encoding' is taken as the type of encoding desired for storing the data in the file. Two
example specifications with UTF-8 and ‘UTF-16’ encoding are given below:
FILE *fp = fopen("myfile.dat", "r+, ccs=UTF-8");
FILE *fp = fopen("myfile.dat", "a, ccs=UTF-16LE");

12.3 CLOSING A FILE

A file must be closed when all operations on the file have completed. When we write data to a file, the
data may not get immediately written physically to the disk. The data may remain in the memory and
get flushed (written) to the disk on an opportune time (may be a size or time threshold). When a file is
closed, all data related to it get flushed to the disk and the file pointer gets invalidated. Keeping too
many open files may also slow down the system. So, we should avoid keeping files unnecessarily open
and should close them when we are done working with them. We can close an already open file using
the function ‘fclose’. It has the following prototype:
fclose(file-pointer);

212
 FILE MANAGEMENT

Given below is a small example where we open two files, one in read mode and another in write mode.
We then close the two files after their use.
FILE *fp1 = NULL;
FILE *fp2 = NULL;
fp1 = fopen(“File1.txt”, “r”);
fp2 = fopen(“File2.txt”, “w”);
…..
…..
fclose(fp1);
fclose(fp2);

12.4 WRITING-TO AND READING-FROM A TEXTUAL FILE

Before using any of the file operation functions we must open the file using the ‘fopen’ function. The
‘FILE’ pointer returned by the ‘fopen’ function is specified as a parameter to all the file operation
functions to identify the file to work with.

12.4.1 WRITING AND READING SINGLE CHARACTERS

We can use the ‘fputc’ (ANSI) and ‘fputwc’ (Wide Character) functions to write single character to a file.
The functions have the following prototypes:
ANSI → int fputc(int iCharacter, FILE *fp);
Wide Character → wint_t fputwc(wchar_t iCharacter, FILE *fp);

The character identified by ‘iCharacter’ is written at the current file position, which is then automatically
advanced by one character. Both the functions are similar except that the later writes Unicode character
instead of ANSI to the file. The file position is the byte position where the read or write operation will
occur next. The file pointer ‘fp’ identifies the file we are working with.

We wrote the character to the file. But, how do we read the character? We can use the functions ‘fgetc’
(ANSI) or ‘fgetwc’ (Wide Character) for the purpose. The functions have the following prototypes:
ANSI → int fgetc(FILE *fp);
Wide Character → wint_t fgetwc(FILE *fp);

Both the functions return the character at the current file position. The file position is then advanced by
one character after the read. The function ‘fgetwc’ returns Unicode character instead of ANSI. If we are
at the end of the file (i.e. no more data is available to be read) the functions return EOF, and sets the end
of file indicator for the file stream.

Now, let us understand the above functions using a simple program. In the program we will write all the
characters corresponding to ASCII value 33 and higher (till ASCII value 126). We will then reopen the
file, read all the characters from the file and print them to the console.

213
Quick & Indepth C With Data Structures

#include <stdio.h>

int main(void)
{
/* Variable Declaration & Initialization. */
FILE *fp = NULL;
int iCharacter = 0;

do
{
/* Create a new file for writing. */
fp = fopen("Characters", "w");
if (!fp)
{
printf("Failed to create the file.");
break;
}

/* Write all the characters one by one within a loop. */

for (iCharacter = 33; iCharacter < 127; iCharacter++)
{
fputc(iCharacter, fp);
}

/* Close the file. */

fclose(fp);
fp = NULL;

/* Reopen the same file, but now for reading. */

fp = fopen("Characters", "r");
if (!fp)
{
printf("Failed to open the file.");
break;
}

printf("The characters are: ");

do
{
iCharacter = fgetc(fp);
if (EOF != iCharacter)
{
printf("%c, ", iCharacter);
}

} while (EOF != iCharacter);

} while (0);

if (fp)
{
fclose(fp);
fp = NULL;
}

printf("\n\n");
return 0;
}

At first glance the above program seems to be doing something funny. We have used a do-while loop
where the loop's exit criteria is '0 != 0' [while (0)]. This means that the condition will never get fulfilled

214
 FILE MANAGEMENT

and the loop will not go into an iteration. So, the body of the loop will always get executed only once
(being a do-while construct). So, why use a loop? We are using the loop just to enable us to stop the
execution at any moment when some condition is not fulfilled. This achieves somewhat similar
functionality of 'goto' statements eliminating some of its disadvantages. This also gets rid of multiple
nested 'if-else' constructs which reduce program readability. Consider the 'if' block starting with 'if (!
fp)' within the loop body. This means, if the file pointer is NULL i.e. the file failed to open, we do not
continue with the program, break out of the loop and exit the function.

12.4.2 WRITING AND READING INTEGERS

The previous functions we studied enable us to write and read characters. For writing an integer to a file
we may use the function ‘putw’ which has the following prototype:
int putw(int iValue, FILE *fp);

The function writes an integer value specified in ‘iValue’ to the file identified by ‘fp’. After writing, the
function advances the file position past the written integer value.

For reading an integer value, we may use the function ‘getw’ which has the following prototype:
int getw(FILE *fp);

The function reads an integer value from the file and returns it. After reading the value, the function
advances the file position past the integer value that was read. If the function reaches end of file, it
returns EOF.

12.4.3 WRITING AND READING STRINGS

If we need to write entire strings to a file, it is inconvenient to be doing so using ‘putc’ within a loop
construct. We need a function which accepts an entire string from us and writes that to the file. This
simplifies our programming by getting rid of unnecessary loop constructs. For writing an entire string to
a file we may use the ‘fputs’ (ANSI) or ‘fputws’ (Wide Character) functions.
ANSI → int fputs(const char *pStr, FILE *fp);
Wide Character → int fputws(const wchar_t *pStr, FILE *fp);

The functions copy from the address pointed by 'pStr' until they reach the terminating NULL character
('\0'). Both the functions are identical in behavior except that ‘fputws’ writes a Unicode string to the
file. The terminating NULL character is not written to the file. The file position is advanced by the
number of characters written to the file. In case of an error the functions return EOF.

To read entire strings from a file, we may use the functions ‘fgets’ (ANSI) or ‘fgetws’ (Wide Character).
Both the functions keep reading the file until end of the specified buffer, a newline character or end of
file is reached. The functions have the following prototypes:
ANSI → char * fgets(char *pStr, int iNumChars, FILE *fp);
Wide Character → wchar_t * fgetws(wchar_t *pStr, int iNumChars, FILE *fp);

215
Quick & Indepth C With Data Structures

We provide a buffer (an array) to both the functions which is then filled with the read string. We must
make sure that the array is large enough to hold an entire string to be read from the file. We pass the
address to the starting location of the array (we may pass the array name directly or the address of the
first element of the array. Please refer to the chapter on pointers). The only difference between the two
functions is that the ANSI version accepts a ‘char’ buffer whereas the wide character version accepts a
‘wide character’ buffer. The ANSI version of the function reads ANSI characters (1 byte each) from the
file and fills the provided buffer with the read string. The wide character version reads wide characters
(1 to 4 bytes each depending on the architecture and file open mode) from the file and fills the provided
buffer with the wide character string.

We pass the size of the array as ‘iNumChars’ to the functions. The functions read a maximum of
‘iNumChars – 1’ characters from the file. So, the functions keep reading characters from the file until
‘iNumChars - 1’ characters have been read or either a newline or the end-of-file is reached, whichever
happens first. The functions then append a NULL character at the end to mark the end of the string. An
important thing to remember is that if both the functions encounter a newline character while reading,
they retain the newline character as the last character of the string (followed by a NULL character).

If the functions encounter an error while reading, they return NULL and the provided buffer remains
unchanged.

12.4.4 WRITING AND READING FORMATTED DATA

Till now we learned file handling functions which read and write only specific type of data viz. Character,
integer or strings. It would make our life far easy if we have functions which help us write and read
multiple items at a time and that too belonging to different data types. We have the functions ‘fprintf’
and ‘fscanf’ which help us to do just that. ‘fprintf’ is very similar to the ‘printf’ function and ‘fscanf’ is
very similar to the ‘scanf’ function. The ‘printf’ function generates the output to the default output
device whereas you can specify the output stream to the ‘fprintf’ function which generally is a file. So,
whatever you write using ‘fprintf’ goes to the specified stream which may be a file handle. Similarly, the
‘scanf’ function reads the input from standard input device, whereas you can specify the input stream in
case of the ‘fscanf’ function. The function will read the input from the specified stream, format the input
and fill the variables with the formatted input data. Let us start our discussion with ‘fprintf’ function.
ANSI → int fprintf(FILE *fp, const char *pszFormat, ...);
Wide Character → int fwprintf(FILE *fp, const wchar_t *pszFormat, ...);

We can see that the function is very similar to the ‘printf’ function except that it accepts an additional
parameter in the form of the file stream identifier. The format specification too is similar to the ‘printf’
function and the function accepts variable number of arguments. An example of its usage is:
fprintf(fp, “The student details are: %s, %d, %f”, szName, iRoll, 81.5);

The function returns the total number of characters written to the file. The function returns a negative
value when it encounters an error.

The ‘fscanf’ function has the following prototype:

216
 FILE MANAGEMENT

ANSI → int fscanf(FILE *fp, const char *pszFormat, argument-list);

Wide Character → int fwscanf(FILE *fp, const wchar_t *pszFormat, argument-list);

The functions read data from the stream (given as 'fp') and store them according to the parameter
format into the locations pointed by the additional arguments (following the parameter 'pszFormat').
The additional arguments should point to variables of the type specified by their corresponding format
specifier within the format string.

We can see that the function is very similar to the ‘scanf’ function except that it accepts an additional
parameter, which is the file stream identifier. The format specification too is similar to the ‘scanf’
function. An example of its usage is:
fscanf(fp, “%d %s”, &iItemCode, szItemName);
where ‘iItemCode’ is an existing integer variable and ‘szItemName’ is a character array.

The function returns the total number of items successfully read and formatted to the specified
arguments list. This may be equal to the expected number of items or be less (even zero) due to a
matching (format) failure, a read error, or on reaching the end of file. The function returns EOF when it
encounters an error.

To understand the above functions, we will write a program which accepts the student name, address,
registration number and class from the user. The program then writes these data to a file.

#include <stdio.h>

/* Defined Values. */
#define MAX_STUDENTS 10
#define ARRAYSIZE(x) sizeof(x)/sizeof(x[0])

/* Structure Definitions. */
typedef struct _STUDENTDETAILS
{
char szName[50];
char szAddress[100];
char szRegNo[20];
unsigned int uClass;

}STUDENTDETAILS;

/* Forward Declarations. */
int AcceptDetails(STUDENTDETAILS *pDetails);
int SaveDetails(FILE *fp, STUDENTDETAILS *pDetails);

int main(void)
{
STUDENTDETAILS Details = { 0 };
int i = 0;
int iReturn = 0;
FILE *fp = NULL;

fp = fopen("Student.dat", "w");
if (!fp)
{
iReturn = -1;

217
Quick & Indepth C With Data Structures

goto EXIT;
}

for (; i < MAX_STUDENTS; i++)

{
iReturn = AcceptDetails(&Details);
if (0 > iReturn)
{
break;
}

iReturn = SaveDetails(fp, &Details);

if (0 > iReturn)
{
break;
}
}

EXIT:
if (fp)
{
fclose(fp);
fp = NULL;
}

return iReturn;
}

int AcceptDetails(STUDENTDETAILS *pDetails)

{
size_t cchLen = 0;

if (!pDetails)
{
return -1;
}

memset(pDetails, 0, sizeof(STUDENTDETAILS));

/* Accept the student's name from the user. The 'fgets' function accepts a string
from the specified stream, which in our case is the standard input device 'stdin'.
'fgets' function keeps reading characters from the input stream until a newline
character ('\n'), end of file (EOF) or the end of the buffer is found. The
function NULL terminates the string but also retains the newline character. As
we do not require the newline character to be present in our string, we overwrite
it with a '\0' character making it as the new end of the string. We have used
'fgets' instead of 'scanf' function because the 'scanf' function stops reading
from the input as soon as it encounters a whitespace character. This makes 'scanf'
unviable for accepting strings like name and address which may contain spaces. */
printf("\nEnter the student's name: ");
fgets(pDetails->szName, ARRAYSIZE(pDetails->szName), stdin);
cchLen = strlen(pDetails->szName);

if (cchLen > 0 && '\n' == pDetails->szName[cchLen – 1])

pDetails->szName[cchLen - 1] = '\0';

printf("Enter the student's address: ");

fgets(pDetails->szAddress, ARRAYSIZE(pDetails->szAddress), stdin);
cchLen = strlen(pDetails->szAddress);

if (cchLen > 0 && '\n' == pDetails->szAddress[cchLen – 1])

pDetails->szAddress[cchLen - 1] = '\0';

218
 FILE MANAGEMENT

printf("Enter the registration number: ");

scanf("%s", pDetails->szRegNo);

printf("Enter the class: ");

scanf("%u", &pDetails->uClass);

return 0;
}

int SaveDetails(FILE *fp, STUDENTDETAILS *pDetails)

{
if (!fp || !pDetails)
{
return -1;
}

fprintf(fp, "%s\n%s\n%s\n%u\n",
pDetails->szName,
pDetails->szAddress,
pDetails->szRegNo,
pDetails->uClass);

return 0;
}

We have accepted and saved the details to a file. Let us now read the saved data and show it to the user.

int LoadDetails(FILE *fp, STUDENTDETAILS *pDetails)

{
int iReturn = -1;
size_t cchLen = 0;

if (!fp || !pDetails)
{
return -1;
}

memset(pDetails, 0, sizeof(STUDENTDETAILS));

/* 'fgets' function keeps reading characters from the file until a newline
character ('\n'), end of file (EOF) or the end of the buffer occurs. The
function retains the newline character. As we do not require the newline
character to be present in our string, we overwrite it with a '\0' character
making it as the new end of the string. We have used 'fgets' instead of
'fscanf' function because the 'fscanf' function stops reading from the input
as soon as it encounters a whitespace character. This makes 'fscanf' nonviable
for accepting strings like name and address which may contain spaces. */

if (NULL == fgets(pDetails->szName, ARRAYSIZE(pDetails->szName), fp))

{
goto EXIT;
}
cchLen = strlen(pDetails->szName);

if (cchLen > 0 && '\n' == pDetails->szName[cchLen – 1])

pDetails->szName[cchLen - 1] = '\0';

if (NULL == fgets(pDetails->szAddress, ARRAYSIZE(pDetails->szAddress), fp))

{
goto EXIT;
}
cchLen = strlen(pDetails->szAddress);

219
Quick & Indepth C With Data Structures

if (cchLen > 0 && '\n' == pDetails->szAddress[cchLen – 1])

pDetails->szAddress[cchLen - 1] = '\0';

if (0 >= fscanf(fp, "%s", pDetails->szRegNo))

{
goto EXIT;
}

/* When reading an integer value, 'fscanf' does not read the newline character.
As a result, the newline character will be read during the next read
iteration upon which 'szName' will contain only the newline character.
To prevent this from happening, we ask the 'fscanf' function to read a
single character after reading the class (uClass). As we have used '*'
in this format specifier '%*c', the character is read but ignored and
dropped. This essentially forwards the read pointer past the newline. */
if (0 >= fscanf(fp, "%u%*c", &pDetails->uClass))
{
goto EXIT;
}

iReturn = 0;

EXIT:
return iReturn;
}

int PrintDetails(STUDENTDETAILS *pDetails)

{
if (!pDetails)
{
return -1;
}

printf("Student Name: %s\n", pDetails->szName);

printf("Student Address: %s\n", pDetails->szAddress);
printf("Student Registration Number: %s\n", pDetails->szRegNo);
printf("Student Class: %u\n\n", pDetails->uClass);

return 0;
}

int main(void)
{
STUDENTDETAILS Details = { 0 };
FILE *fp = NULL;

fp = fopen("Student.dat", "r");
if (!fp)
{
return -1;
}

while (LoadDetails(fp, &Details) >= 0)

{
PrintDetails(&Details);
}

fclose(fp);
fp = NULL;

return 0;
}

220
 FILE MANAGEMENT

12.5 WRITING-TO AND READING-FROM A BINARY FILE

Till now we focused on writing and reading in textual format. The biggest disadvantage of using textual
format is that we have to write each individual items separately. Moreover, we have to take care of the
encoding of the text we intend to write or read. If we are working with data types like integer or float,
we need to use formatted output and input for writing and reading data. This involves additional
processing overheads as the data is converted to string and back. To overcome these disadvantages, C
provides us with two functions 'fwrite' and 'fread' which enable us to write and read data as is i.e. the
way they are represented in memory.

When we use functions like 'fwrite' and 'fread', we need to open the files in binary format mode. We have
to specify the additional flag 'b' in the 'mode' parameter of 'fopen' function (refer to the section on
'fopen'). When using 'fwrite' on integer or real values, the data is not converted to textual format but are
written as they are represented in memory. Hence, when reading the data back using 'fread', the data
does not require to be converted back.

The biggest advantage of the functions 'fwrite' and 'fread' is that they allow writing and reading data in
bulk. For 'fwrite', it is like dumping the entire memory representation of the data directly to disk, and
for 'fread', it is like retrieving the entire dumped data from the disk to the memory. For this reason,
these two functions are much faster when operating on large amounts of data than their textual
counterparts.

Writing and reading data in bulk reduces coding time for data types like structures and unions. Rather
than writing each individual item (member or field) of the structure separately, we may write the entire
structure data directly at one go using 'fwrite'. Similarly, we may read the entire structure data directly
using a single call to 'fread'.

Now, let us have a look at the prototypes of the two functions starting with 'fwrite':
size_t fwrite(const void *pData, size_t ItemSize, size_t ItemCount, FILE *fp);

The function writes an array of 'ItemCount' elements, each with a size of 'ItemSize' bytes, from the
memory location pointed by 'pData' to the specified file (identified using 'fp'). The data is written at the
current position within the file. For example, if we are writing an array of structure data, we need to
specify the array name (or the address of the first element of the array) as 'pData'. We will specify the
size of each array element (equivalent to the structure size) as 'ItemSize'. The number of array elements
we intend to write should be specified for 'ItemCount'.

In place of writing the entire array data at one go, we may utilize a loop within which we will write each
array element using the 'fwrite' function. Each iteration of the loop will write a single array element by
specifying the element address as 'pData', the element size as 'ItemSize' and 1 (one) as 'ItemCount'. In
the next iteration, we will increment to point to the next array element.

After writing the data, position indicator of the file is automatically advanced by the total number of

221
Quick & Indepth C With Data Structures

bytes written. The function returns the total number of elements (items) written to the file. If the
number of elements written is different from 'ItemCount', we can assume that there was an error which
prevented the function from writing all the data.

Now, let us have a look at the 'fread' function:

size_t fread(void *pData, size_t ItemSize, size_t ItemCount, FILE *fp);

The function reads an array of 'ItemCount' elements, each with a size of 'ItemSize' bytes, to the memory
location pointed by 'pData' from the specified file (identified using 'fp'). The data is read from the
current position within the file. We must make sure that the memory location pointed by 'pData' is large
enough to hold the read data, which typically must be at-least of (ItemSize * ItemCount) bytes.

After reading the data, position indicator of the file is automatically advanced by the total number of
bytes read. The function returns the total number of elements (items) read from the file. If the number
of elements read is different from 'ItemCount', we can assume that there was an error which prevented
the function from reading all the data.

We will understand the behavior of the two functions better using two programs. In both the programs
we will accept details of items used in a consumer shop. Each item detail will be stored in a single
structure variable.

In our first program, we will accept each item's detail from the user and immediately save that data to a
file. This will be done in each iteration. After accepting and saving the details, we will start reading the
data back from the file. We will read each item individually and print its details. This approach is
convenient in situations where we are not sure regarding the count of items we are going to save to the
file or read from the file.

In our second program, we will declare an array of the structure and will accept the details of all the
items in the array. We will then save the entire array at one go to the file. When reading the data back
from the file, we will read the entire data to our array at one go.

PROGRAM-1
#include <stdio.h>
#include <ctype.h>

/* Structure Definition. */
typedef struct _ITEMDETAILS
{
int iItemCode;
char szItemName[50];
float fltItemRate;

}ITEMDETAILS;

/* Function which accepts the details from the user. */

int AcceptDetails(ITEMDETAILS *pDetails)
{
int iReturn = 0;

222
 FILE MANAGEMENT

do
{
if (!pDetails)
{
break;
}

printf("\nEnter the item code: ");

scanf("%d", &pDetails->iItemCode);

printf("Enter the item name: ");

scanf("%s", pDetails->szItemName);

printf("Enter the item rate: ");

scanf("%f", &pDetails->fltItemRate);

iReturn = 1;

} while (0);

return iReturn;
}

/* Function for printing the item details to the console. */

int PrintDetails(ITEMDETAILS *pDetails)
{
int iReturn = 0;

do
{
if (!pDetails)
{
break;
}

printf("\n\nItem code: %d", pDetails->iItemCode);

printf("\nItem name: %s", pDetails->szItemName);
printf("\nItem rate: %f", pDetails->fltItemRate);

iReturn = 1;

} while (0);

return iReturn;
}

int main()
{
/* Variable Declaration and Initialization. */
ITEMDETAILS ItemDetails = { 0 };
char chContinue = 'y';
size_t iReturn = 0;
FILE *fp = NULL;

/* Open the file for writing. */

fp = fopen("Items.dat", "wb");
if (!fp)
{
/* Failed to create the file. Bail out with a message.*/
printf("Failed to create the file. Stopping...");
goto EXIT;
}

223
Quick & Indepth C With Data Structures

do
{
/* Accept the details from the user in our structure variable.*/
AcceptDetails(&ItemDetails);

/* Write the entire structure contents directly to the file at

the current file position which will be advanced after the write. */
iReturn = fwrite(&ItemDetails, sizeof(ItemDetails), 1, fp);
if (1 != iReturn)
{
printf("\nError writing to file. Continuing to read...");
break;
}

printf("Press Y/y to continue: ");

chContinue = getchar();

} while ('y' == tolower(chContinue));

/* Close the file and reopen it for reading. */

fclose(fp);
fp = NULL;

fp = fopen("Items.dat", "rb");
if (!fp)
{
/* Failed to open the file. Bail out with a message.*/
printf("Failed to open the file for reading...");
goto EXIT;
}

/* Keep reading until the end of file is reached. */

while (1 == fread(&ItemDetails, sizeof(ItemDetails), 1, fp))
{
/* Print the details. */
PrintDetails(&ItemDetails);
}

EXIT:
if (fp)
{
fclose(fp);
fp = NULL;
}

printf("\n\n");
return 0;
}

PROGRAM-2
For this program, we will have identical 'ITEMDETAILS' structure definition as Program-1. We will also
have identical 'AcceptDetails' and 'PrintDetails' function definition as Program-1. So, we are not re-
specifying those over here and are just concentrating on the 'main()' function which differs between the
two.
#define MAX_ITEMS 10

int main()
{

224
 FILE MANAGEMENT

/* Variable Declaration and Initialization. */

ITEMDETAILS ItemDetails[MAX_ITEMS] = { 0 };
int iCount = 0;
size_t iReturn = 0;
FILE *fp = NULL;

/* Open the file for writing. */

fp = fopen("Items.dat", "wb");
if (!fp)
{
/* Failed to create the file. Bail out with a message.*/
printf("Failed to create the file. Stopping...");
goto EXIT;
}

/* Accept the details from the user in our structure array.*/

for (iCount = 0; iCount < MAX_ITEMS; iCount++)
{
AcceptDetails(&ItemDetails[iCount]);
}

/* Write the details to the file at one go. */

iReturn = fwrite(ItemDetails, sizeof(ItemDetails[0]), iCount, fp);
if (iReturn != iCount)
{
/* Failed to write the data to the file. Bail out...*/
printf("\nFailed to write the data to the file. Stopping...");
goto EXIT;
}

/* Close the file and reopen it for reading. */

fclose(fp);
fp = NULL;

fp = fopen("Items.dat", "rb");
if (!fp)
{
/* Failed to open the file. Bail out with a message.*/
printf("Failed to open the file for reading...");
goto EXIT;
}

memset(ItemDetails, 0, sizeof(ItemDetails));

/* Read the entire file at one go. */

iReturn = fread(ItemDetails, sizeof(ItemDetails[0]), iCount, fp);
if (iReturn != iCount)
{
/* Failed to read the data from the file. Bail out...*/
printf("\nFailed to read the data from the file. Stopping...");
goto EXIT;
}

/* Print all the read items. */

for (iCount = 0; iCount < iReturn; iCount++)
{
/* Print the details. */
PrintDetails(&ItemDetails[iCount]);
}

EXIT:
if (fp)
{

225
Quick & Indepth C With Data Structures

fclose(fp);
fp = NULL;
}

printf("\n\n");
return 0;
}

12.6 MISCELLANEOUS FILE MANAGEMENT FUNCTIONS

Apart from the input and output functions, C has some additional functions which help in working with
files.

• int fgetpos(FILE *fp, fpos_t *pos);

The function retrieves the current byte position within the file (identified by 'fp'). This is the
position where the next read or write operation will occur, if not changed. The variable 'pos'
needs to be declared by the caller and is filled with the current position after the call completes.
'fpos_t' is compiler defined and may be a type-defined name for 'long' or 'long long'. The
function returns 0 (zero) in case of success and a non-zero value in case of an error.

• int fsetpos(FILE *fp, const fpos_t *pos);

The function sets the current byte position to 'pos'. This is the position where the next read or
write operation will occur. The variable 'pos' needs to be declared and set by the caller. It may
also be filled using a previous call to 'fgetpos'. The function returns 0 (zero) in case of success
and a non-zero value in case of an error.

• long int ftell(FILE *fp);

The function returns the current byte position within the file (identified by 'fp'). It is very similar
to 'fgetpos' except that it returns a 'long' value instead of 'fpos_t' for 'fgetpos'. So, this function
will fail if the file size is larger than 2 GB. There are compiler specific functions which overcome
this limit (ex: '_ftelli64' in Microsoft VC or 'ftello' in POSIX). In case of an error, this function
returns -1.

• int fseek(FILE *fp, long int lOffset, int iOrigin);

The function sets the current byte position of the file (identified by 'fp'). 'iOffset' is the number
of bytes of offset from 'iOrigin'. When 'iOffset' is positive, the file position moves forward.
'iOffset' can be negative, implying that the file position be moved backwards as respect to
'iOrigin'. 'iOrigin' may have the following compiler-defined values:
SEEK_SET: As respect to beginning of the file.
SEEK_CUR: As respect to current byte position within the file.
SEEK_END: As respect to the end of the file.
As the function accepts a 'long' value as 'iOffset', we are limited to operating with 2 GB limit.
There are compiler specific functions which overcome this limit (ex: '_fseeki64' in Microsoft VC
or 'fseeko' in POSIX). The function returns 0 (zero) if successful, and a non-zero value in case of
an error.

226
 FILE MANAGEMENT

• void rewind(FILE *fp);

Sets the byte position to the beginning of the file.

• FILE * freopen(const char *pszFilename, const char *pszMode, FILE *fp);

The function reopens the file stream with different file or mode. If a new filename is specified as
'pszFilename', the function closes the existing file. The function then opens the file specified by
'pszFilename' using the specified mode 'pszMode', and associates it with the file stream handle
'fp', just like 'fopen' would do. If 'pszFilename' is specified as NULL, the function attempts to
change the file open mode to 'pszMode'. If the function succeeds, it returns the same pointer
'fp' specified as the third parameter to the function. If the function fails, it returns NULL.

• int remove(const char *pszFilename);

Deletes the file whose name is specified as 'pszFilename'. The operation is performed directly on
the file identified by 'pszFilename'. 'pszFilename' may be an absolute path to a file. If only the
name of the file is specified, it is assumed that the file resides in the current working directory.
The file does not need to be opened for the file to be deleted. The function returns 0 (zero) if
the file was deleted, and a non-zero value in case of an error.

• int rename(const char *pszOldname, const char *pszNewname);

The function renames the file or directory specified by 'pszOldname' to the name specified by
'pszNewname'. The old name must be the path of an existing file or directory and the new name
must not be an existing file or directory. We can also use this function to move a file from one
directory to another by giving a different path in the 'pszNewname' argument. However, the
function cannot move a directory.

• FILE * tmpfile(void);
The function creates a temporary file in 'wb+' mode (refer to 'fopen') with a filename
guaranteed to be different from any other existing file. The file is created in the current working
directory. The file is automatically deleted when the file is closed using 'fclose' or the program
terminates. If the function fails, the return value is NULL.

• char * tmpnam(char *pszStr);

The function returns a string containing a filename different from the name of any existing file.
This enables us to safely create a temporary file without the risk of overwriting an existing file.
If 'pszStr' is NULL, 'tmpnam' leaves the result in an internal static buffer. Thus any subsequent
calls to 'tmpnam' will destroy this value. It is recommended that the caller provides a buffer
(char array) for 'pszStr'. The function will then populate this buffer with the generated
temporary filename. The buffer must be large enough to hold the temporary filename.

227
Quick & Indepth C With Data Structures

DYNAMIC MEMORY ALLOCATION

[CHAPTER-13]

13.1 INTRODUCTION
During programming, many a times we may face situations where we are unsure regarding the amount
of data we are going to deal with. The amount of data is determined during execution of the program,
and so we are completely in the dark when coding. Consider an example where we are accepting marks
obtained by students of different classes and calculating the rank and total percentage obtained by
each student. Not only the number of students change, but also the number of subjects change
depending on the class. How do we provision the memory for such dynamic nature of data? For such
situations, we have what we call dynamic memory allocation. As the name suggests, in this mechanism
the memory allocation is entirely handled dynamically during execution of the program. Dynamic
memory allocation or dynamic memory management allow us to allocate additional memory as per our
requirement and also free those when they are not required. Dynamic memory allocation is extensively
used in concepts like linked lists and trees which we will learn later in our book.

13.2 C PROGRAM – MEMORY LAYOUT

Before going deeper into dynamic memory allocation, let us understand the memory layout of a C
program.

228
 DYNAMIC MEMORY ALLOCATION

A C program contains many components such as the program code, function local variables, global
variables, static variables, dynamically allocated memory space etc. All these need to be provided with
some memory space to operate. The processor pulls these data from the memory rather than the disk.
So, all these components need to be present in the memory for the CPU to work on them. A program
may be saved to disk, but is loaded to memory before it is run, and the entire execution operates from
the memory itself.

Code or Text segment:

This section contains the executable instructions. It is the machine code of the compiled program. This
segment is mostly read-only as it is not allowed to be accessed by the program code. This segment is
shareable so that only a single copy needs to be in memory for frequently executed programs, such as
text editors etc. Only one shareable segment optimizes memory utilization of the operating system.

Initialized data segment:

Initialized data segment contains the global variables and static variables that are initialized by the
programmer within the program. This segment may contain two types of data – initialized read-only
data or initialized read-write data. Read-only data are the ones which are initialized during their
declaration and cannot be changed thereafter, such as a constant global variable. Read-write data are
the ones which can be altered any time during the program execution, such as normal global variables.

Uninitialized data segment:

Uninitialized data segment, also known as the 'bss' segment, is named after an ancient assembler
operator that stood for 'block started by symbol'. Data in this segment is automatically initialized to 0
(zero) by the operating system before the program starts. This segment contains all global and static
variables that have not been explicitly initialized in the source code by the programmer. As for an
example, a static variable declared as ‘static int iValue;’ or a global variable declared as ‘float g_fltRate;’
will be put in this segment.

Stack segment:
This is the segment where automatic variables are stored. Also, each time a function is called, the
address (next instruction location within the code segment) of where to return to (after the function
completes) is saved in this segment. Assume, function 'f1' calls the function 'f2'. The next instruction
location is within the caller function 'f1' and is the instruction which needs to be executed after the
function (f2) call completes. The newly called function (f2) then allocates room on the stack for storing
its automatic and temporary variables. This is how recursive functions also work. Each time a recursive
function calls itself, a new stack space is used for the new call. So one set of variables belonging to one
instance does not interfere with the variables from another recursive instance of the function.

Heap segment:
Heap is the segment which stores the data that are dynamically allocated. All dynamically allocated data
goes into this segment. We use the C functions ‘malloc’, ‘calloc’ and ‘realloc’ to allocate memory
dynamically from this segment. We use the C function ‘free’ to free the dynamically allocated space
which are no longer required. When using the allocation functions, we specify the amount of memory (in

229
Quick & Indepth C With Data Structures

bytes) we require, and consecutive memory location from this segment is allotted to us. After the
memory is allotted, the starting address (memory location) of the allotted region is returned to us. So,
essentially we declare a pointer which points to this starting address. Unlike stack segment, the data
within heap segment may be shared between functions of the program, shared libraries or even
dynamically loaded modules of the program. Just like stack segment, data in this segment is
uninitialized and may contain garbage value after being allotted.

13.3 ALLOCATING MEMORY

We may allocate memory using the C function ‘malloc’. The function reserves a contiguous block of
memory from the heap segment and returns the starting address of the block. The function needs to be
specified the number of bytes (memory-size) to be be allocated and the function returns a void pointer
pointing to the starting location of the allocated memory. Returning a void pointer allows us to typecast
it to a pointer of any type. So, the function returns only the block of requested memory space. What we
store in that memory space (int, float, char, structure, string...) depends on us, provided the allocated
space is equal or larger than the data we are trying to store. The function has the following form:
void* malloc(size_t SizeInBytes);

The returned void pointer can be type-casted to a pointer of any desired type. If the function fails to
allocate the requested space, it returns a NULL pointer. Few example usages of the function are:
/* Allocate space for a single double variable */
double *pdblRate = NULL;

pdblRate = (double *) malloc(sizeof(double));

*pdblRate = 62872.7372;
/* Allocate space for an array of integers */
int iCounter = 0;
int *prgiValues = NULL;

prgiValues = (int *) malloc(100 * sizeof(int));

for (iCounter = 0; iCounter < 100; iCounter++)
{
prgiValues[iCounter] = iCounter;
}
/* Allocate space to store a character string */
char *pszName = NULL;

pszName = (char *) malloc(13 * sizeof(char));

strcpy(pszName, “David Miller”);
/* Allocate space for an array of structures */
int iCounter = 0;
struct ITEMDETAILS *pDetails = NULL;

230
 DYNAMIC MEMORY ALLOCATION

pDetails = (struct ITEMDETAILS *) malloc(sizeof(struct ITEMDETAILS) * 100);

for (iCounter = 0; iCounter < 100; iCounter++)
{
pDetails[iCounter].iItemID = iCounter;
}
/* Allocate space for an array of structures */
int iCounter = 0;
struct ITEMDETAILS *pDetails = NULL;
struct ITEMDETAILS *pMove = NULL;

pDetails = (struct ITEMDETAILS *) malloc(sizeof(struct ITEMDETAILS) * 100);

for (iCounter = 0, pMove = pDetails; iCounter < 100; iCounter++, pMove++)
{
pMove->iItemID = iCounter;
}

When we allocate a space and assign it to a pointer, the pointer points to the starting address of the
allocated space.

Please note that the allocated memory space initially contains garbage data and we should avoid
accessing the value of any location before initializing it. We should also be careful regarding overflowing
our allocated memory space i.e. trying to access a memory location outside of what was allotted to us.
Please note that we cannot determine the size of the reserved memory space by using the ‘sizeof’
operator on the returned memory pointer. Using the ‘sizeof’ operator on the pointer will return the
memory space used by the pointer variable itself and not the memory space pointed by it. This will
typically be 4 bytes on a 32-bit architecture and 8 bytes on a 64-bit architecture. In our above image,
the returned (sizeof) value will be the space used by the ‘ptr’ variable itself and not the memory space
pointed by the ‘ptr’ pointer variable.

C provides another memory allocation function called ‘calloc’ which allows allocating memory space at
run time. It is similar to ‘malloc’ except for the following two differences:
1. When allocating an array, we need to calculate the total space required by the array (number of
array elements multiplied by the size of each element) and specify that to ‘malloc’. In case of
‘calloc’, we do not require to do this calculation. We specify the number of elements of the array
and the size of each element as two different parameters to the function ‘calloc’. The function
calculates the size required before doing the allocation.
2. In case of ‘malloc’, the newly allocated space is uninitialized and contains garbage data. ‘calloc’

231
Quick & Indepth C With Data Structures

initializes the newly allocated space to 0 (zero). So, every byte of the allocated memory space
will contain the value 0 (zero) after allocation by ‘calloc’.

The general form of the ‘calloc’ function is:

void* calloc(size_t NumElements, size_t SizeOfEachElement);

Allocates a block of memory for an array of ‘NumElements’ elements, each of ‘SizeOfEachElement’ bytes
long, and initializes all its bytes to 0 (zero). So, effectively it allocates a memory block of (NumElements
* SizeOfEachElement) bytes.

The function returns a pointer to the starting address of the allocated memory space which can be
type-casted to any desired type. If the function fails, it returns a NULL pointer.

13.4 ALTERING ALLOCATED MEMORY SIZE

It is very likely that later in our program we discover that the previously allocated memory is not
sufficient to hold all our data (or array elements) and we need additional space for more data (or array
elements). It is also possible that we later realize that we had been over ambitious and allocated too
much space, much more than what we require. In both these cases, we can change the already
allocated memory size using the help of another function called ‘realloc’. The function name is the short
form of reallocation of memory. The general form of the function is:
void* realloc(void *ptr, size_t NewSizeInBytes);

The function changes the size of the memory block pointed to by 'ptr' to the new size 'NewSizeInBytes'.
The function may move the memory block to a new location (whose address is returned by the function).
If the new requested memory space is larger than the previously allocated space, the contents of the
old memory space is entirely copied to the new location. If the new requested memory space is smaller,
the contents of the old memory space up to a maximum of the newly requested space is copied to the
new location. Hence in simple words, the content of the memory block is preserved up to the lesser of
the old and new sizes.

The function returns a pointer pointing to the address of the starting location of the newly allocated
memory block. The new memory block may or may not begin at the same location as the old block. If
the function does not find enough additional contiguous space in the same region, it will allocate a
newly requested space in another location and return that address. The function also frees the old
memory location after the transfer to the new location. If it is able to allocate the space in the old
location itself, it will return the same address as ‘ptr’. An example usage of the ‘realloc’ function is:
void *ptr = malloc(50 * sizeof(int));
ptr = realloc(ptr, 100 * sizeof(int));

13.5 RELEASING ALLOCATED SPACE

Dynamically allocated space need to be freed by the programmer when it is no longer required. We
should not keep holding onto memory space when not needed, as memory space is limited. When we

232
 DYNAMIC MEMORY ALLOCATION

release a memory block, it becomes available again for allocating. We free an already allocated memory
space using the function ‘free’. Memory allocated using ‘malloc’, ‘calloc’ or reallocated using ‘realloc’
can be freed using this function. The general form of the function is:
void free(void *ptr);

‘ptr’ is the pointer pointing to the memory block allocated using the functions ‘malloc’, ‘calloc’ or
reallocated using the function ‘realloc’. Using an invalid address in ‘ptr’ may result in undesirable
effects.

13.6 SAMPLE PROGRAM

Let us write a program which accepts the details of the books in a library. For simplicity, we will accept
only two details related to the books – their IDs and their names. We will use a structure to hold the
information regarding each book. We will hold all the accepted information using an array of the
structures in memory. Each element of the array will have details regarding one book. As we do not
know beforehand regarding the number of book details the user wants to enter, we will allocate memory
as required during run time. If our already allocated space is not large enough to store any more data,
we will reallocate the memory for a larger space. After we have accepted all the details, we will print
those.

#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>

/* Defined Values */
#define SIZE_INC_COUNT 10

/* Defined Structures */
typedef struct _BOOKDETAILS
{
int iBookID;
char szBookName[100];

}BOOKDETAILS, *PBOOKDETAILS;

int main(void)
{
/* Variable Declaration & Initialization. */
PBOOKDETAILS pBooks = NULL; /* Same as stating BOOKDETAILS *pBooks */
PBOOKDETAILS pMove = NULL;
int iCurArraySize = 0;
int idxInput = 0;
char chContinue = 'y';

/* Keep accepting book details till the user chooses to continue.

We will continue till the user presses 'Y' or 'y'. */
for ( ; 'y' == tolower(chContinue); idxInput++)
{
if (idxInput >= iCurArraySize)
{
/* The user wants to continue, but we have run out of array elements.
If we have not yet allocated any memory for the array, we will use
'calloc', and if we need to add to already allocated space, we will
use 'realloc'. 'iCurArraySize' tracks our current allocated array

233
Quick & Indepth C With Data Structures

size which is incremented everytime we request for more space. */

iCurArraySize += SIZE_INC_COUNT;

if (0 == idxInput)
{
/* We have not allocated any memory space yet. So, we use 'calloc'. */
pBooks = (PBOOKDETAILS)calloc(iCurArraySize, sizeof(BOOKDETAILS));
}
else
{
/* We need more space. So, we request a larger memory space. */
pBooks = (PBOOKDETAILS)realloc(pBooks, iCurArraySize * sizeof(BOOKDETAILS));
}

if (NULL == pBooks)
{
printf("\nFailed to allocate sufficient memory. Stopping...");
break;
}
}

printf("Enter the book ID: ");

scanf("%d", &pBooks[idxInput].iBookID);

printf("Enter the book name: ");

fgets(pBooks[idxInput].szBookName, 100, stdin);

printf("\nPress 'Y/y' to continue: ");

chContinue = getchar();

printf("\n");
}

/* We have accepted the data from the user. Let us now print the data.
For accepting the data, we have demonstrated using the array index.
For printing the data, we will use pointers and pointer arithmetic.
We will not increment the 'pBooks' pointer but use a separate pointer
'pMove'. When freeing the allocated memory block, we need to specify
the starting address of the memory block to the function 'free'. As,
'pBooks' points to the starting address, we can use that later in
our program to free the allocated memory block/space. */
if (pBooks != NULL)
{
for (pMove = pBooks; (pMove - pBooks) < idxInput; pMove++)
{
printf("Book ID: %d, Book name: %s", pMove->iBookID, pMove->szBookName);
}

/* We have completed using our allocated memory. Now, free it. */

free(pBooks);
pBooks = NULL;
}

printf("\n");
return 0;
}

13.7 THINGS TO REMEMBER

1. The memory allocation functions 'malloc', 'calloc' and 'realloc' may fail, in which case they will
return NULL. Before trying to use the returned pointer for any operation, we must verify that the

234
 DYNAMIC MEMORY ALLOCATION

pointer is not NULL. Trying to use a NULL pointer will generate a memory fault leading to a
program crash.
2. Never provide an invalid or NULL pointer to the function 'free'. If any of the memory allocation
functions failed and returned NULL, we have nothing to free.
3. Always be careful regarding the amount of memory we have reserved. We should never try to
access a location outside the allocated memory block. Though pointer arithmetic is possible for
locations beyond the memory block (example: incrementing a pointer to a location beyond the
memory block), we should never try to access or manipulate those locations.
4. When freeing a memory block, always provide the address of the starting location of the
memory block to the function 'free'.
5. We should not rely too much on dynamic memory allocation. Too many number of allocations
and deallocations lead to memory fragmentation. This may affect the performance of our
program. Also, when memory becomes fragmented, contiguous memory locations may cease to
exist, leading to memory allocation failures.

235
Quick & Indepth C With Data Structures

COMMAND LINE ARGUMENTS

[CHAPTER-14]

14.1 INTRODUCTION
On many occasions we may want our program to be non-interactive. The program may also need to
work as a background or hidden process and be invoked from other processes for helping them
accomplish certain tasks. If our program needs to work with certain input values, how do other
programs pass those values to us? The answer is command line arguments. Command line arguments
are parameters supplied to the program when it is invoked. They are used when we need to control our
program or pass values to our program from outside. Also, rather than hard-coding our program with
specific values, we may accept the values as command line arguments from the user. When executing
the program, command-line arguments are specified after the program name with each argument
separated from the other by a space. They are passed into the 'main()' function of the program by the
operating system. Examples of invoking a program using command line arguments are:
myprog1 inputfile.dat outputfile.dat
myprog2 http://www.mydomain.com/

14.2 ‘main’ FUNCTION REVISITED

Till now we have been mainly dealing with ‘main’ function accepting nothing (‘void’) as its parameter.
Instead of ‘void’, it may have two specific types of parameters which enable it to accept command line
options for the program. Any command line arguments we specify when invoking the program are
passed to the ‘main’ function as parameters to it. The prototype or declaration of such a ‘main’ function
is as follows:
int main(int argc, char *argv[]);
ANSI → or
int main(int argc, char **argv);
int wmain(int argc, wchar_t *argv[]);
Wide Character → or
int wmain(int argc, wchar_t **argv);

Description of the parameters are:

• argc (argument count): It is an integer value specifying the count of arguments passed to the
program from the command line. The program name is considered as the first argument and
thus, the value of ‘argc’ is at-least 1. Each command line argument following the program name
when invoking (executing) the program increments the value of ‘argc’. So, ‘argc’ will have a
value of 4 for a command line as ‘myprogram value1 value2 value3’.
• argv (argument vector): An array of null-terminated strings. The first string (argv[0]) is the
program name, and each following string (argv[1], argv[2], ...) is an argument passed to the
program from the command line. As the program name is considered as an argument, we will

236
 COMMAND LINE ARGUMENTS

always have the program name for 'argv[0]'. All subsequent strings (argv[1], argv[2], ...)
depend on whether our program was invoked using any additional command line arguments or
not. So, the last valid string will be 'argv[argc – 1]'. The last pointer (argv[argc]) is always
NULL. If the program was invoked as ‘myprogram inputfile.dat outputfile.dat’, ‘argv[0]’ will have
the string ‘myprogram’, ‘argv[1]’ will have the string ‘inputfile.dat’ and ‘argv[2]’ will have the
string ‘outputfile.dat’.

14.3 SAMPLE PROGRAMS

1. In our first program, we will copy one file to another. We will accept two command line
arguments corresponding to two filenames. The first filename is the input file i.e. the file to be
copied. The second filename is the target file i.e. the one which needs to be created. The
program opens the first file, creates the second and copies all the data (byte at a time) to the
second file.

#include <stdio.h>

int CopyFile(char *pszInputFile,

char *pszOutputFile)
{
/* Variable Declaration and Initialization. */
FILE *fp = NULL;
FILE *tp = NULL;
int iReturn = 0;
unsigned char cValue = 0;

do
{
if (!pszInputFile || !pszOutputFile)
{
iReturn = -1;
break;
}

/* Open the input file for reading. */

fp = fopen(pszInputFile, "rb");
if (!fp)
{
iReturn = -2;
break;
}

/* Open the output file for writing. */

tp = fopen(pszOutputFile, "wb");
if (!tp)
{
iReturn = -3;
break;
}

/* Read each byte from the input file and write that to
the output file. Continue, till all the bytes are read. */
while (1 == fread(&cValue, sizeof(cValue), 1, fp))
{
if (1 != fwrite(&cValue, sizeof(cValue), 1, tp))
{

237
Quick & Indepth C With Data Structures

iReturn = -4;
break;
}
}

} while (0);

if (fp)
{
fclose(fp);
fp = NULL;
}

if (tp)
{
fclose(tp);
tp = NULL;
}

return iReturn;
}

int main(int argc, char *argv[])

{
int iReturn = 0;

if (argc != 3)
{
printf("Usage: myprogram <InputFilename> <OutputFilename>\n\n");
return -1;
}

/* Copy the file. */

iReturn = CopyFile(argv[1], argv[2]);
if (iReturn == -2)
{
printf("Error opening input file %s\n\n", argv[1]);
}
else if (iReturn == -3)
{
printf("Error creating output file %s\n\n", argv[2]);
}
else if (iReturn == -4)
{
printf("Error writing to output file %s\n\n", argv[2]);
}
else if (iReturn == 0)
{
printf("File successfully copied.\n\n");
}
else
{
printf("Error copying file.\n\n");
}

return iReturn;
}

2. In our next program we will calculate the compound interest on bank deposits. Compound
interest is that interest which is added back to the principal sum, so that the interest is earned
on that added interest too during the next compounding period. The shorter the compounding

238
 COMMAND LINE ARGUMENTS

period the higher will be the maturity amount. Compounding frequency is the frequency at
which the interest is calculated (ex. every month, every quarter, every twelve months). A monthly
compounding (compound frequency of 12) will earn a higher maturity amount than a half-yearly
compounding (compound frequency of 2), as the interest on interest is calculated much more
frequently. The formula for annual compound interest is:
A = P * (1 + r / n) (n * t)

Where:
A = The amount receivable including interest after the maturity period
P = The principal investment amount (the initial deposit)
r = The annual rate of interest
n = The number of times the interest is compounded per year
t = The period of interest

In the program we will accept the principal investment amount, annual rate of interest,
compounding frequency and the period of investment through the command line. We will then
calculate the total maturity amount for the investment.

/* Header Files */
#include <stdio.h>

/* Defined Structures */
typedef struct _CMPDINT
{
double dblPrincipal;
float fltInterestRate;
int iCompoundFrequency;
float fltPeriod;
double dblMaturityAmount;

}CMPDINT;

void CalculateMaturity(CMPDINT *pcmpdint)

{
/* Variable Declaration & Initialization */
double dblRate = 0;
double dblPeriod = 0;

if (pcmpdint)
{
/* fltInterestRate is in percentage. Convert to decimals. */
dblRate = (double)pcmpdint->fltInterestRate / 100;

dblRate = 1 + (dblRate / pcmpdint->iCompoundFrequency);

dblPeriod = pcmpdint->iCompoundFrequency * (double)pcmpdint->fltPeriod;

pcmpdint->dblMaturityAmount = pcmpdint->dblPrincipal * pow(dblRate, dblPeriod);

}
}

int main(int argc, char *argv[])

{
int iReturn = -1;
CMPDINT cmdint;

239
Quick & Indepth C With Data Structures

if (argc != 5)
{
printf("Usage: myprogram <Principal> <Rate> <Frequency> <Period [Years]>\n\n");
return iReturn;
}

cmdint.dblPrincipal = atof(argv[1]);
cmdint.fltInterestRate = atof(argv[2]);
cmdint.iCompoundFrequency = atoi(argv[3]);
cmdint.fltPeriod = atof(argv[4]);

if (cmdint.dblPrincipal <= 0)
{
printf("Principal amount must be greater than 0.\n\n");
}
else if (cmdint.fltInterestRate <= 0)
{
printf("Interest rate must be greater than 0.\n\n");
}
else if (cmdint.iCompoundFrequency > 12 || cmdint.iCompoundFrequency < 1)
{
/* 12 --> monthly compounding.
4 --> quarterly compounding.
2 --> half yearly compounding.
1 --> yearly compounding. */
printf("Compound frequency must be between 1-12.\n\n");
}
else if (cmdint.fltPeriod <= 0)
{
printf("Period must be greater than 0.\n\n");
}
else
{
CalculateMaturity(&cmdint);
printf("Total maturity amount will be %lf.\n\n", cmdint.dblMaturityAmount);
iReturn = 0;
}

return iReturn;
}

240
 PREPROCESSOR DIRECTIVES

PREPROCESSOR DIRECTIVES

[CHAPTER-15]

15.1 INTRODUCTION
C Preprocessor directives are special statements which are used to make a program portable, easy to
read, easy to modify, efficient and being conditionally compiled. As the name suggests, preprocessor
directives are processed before the source code is passed through the compiler. Before a source code is
sent to the compiler, it is analyzed by the preprocessor for any such statements and if found, the source
code is modified (in memory) depending on the statements and then sent to the compiler. All
preprocessor directives begin with the character ‘#’ which needs to be the first non-blank character in
the line. Also preprocessor directives do not end with a semicolon. We are very accustomed to the
preprocessor directives ‘#define’ and ‘#include’. All preprocessor directives can be divided into three
categories:
• File inclusion directives
• Macro directives
• Compiler control directives

15.2 FILE INCLUSION DIRECTIVES

From the start of our book we have been using statements like ‘#include <stdio.h>’. It is a file inclusion
statement which includes the specified header file during compilation. So, what do statements like these
actually do? Whenever we use the ‘#include’ directive, we are essentially asking the preprocessor to
copy the entire contents of the specified file at the exact location where the ‘#include’ directive is
specified. So, essentially the directive gets replaced by the contents of the file mentioned in the
directive. This makes the source code file contain the contents of the included file. All this is done in
memory before sending the modified source code to the compiler for compilation. The actual source
code file present in the disk isn’t affected in any way.

The included file may further include other files. Included files generally contain function declarations
(system header files contain system library function declarations), type-defined values, defined macros,
structure or union declarations etc. When the included file is specified within double-quotation marks in
the ‘#include’ directive, the file is first searched in the current directory and then in the standard
directories. If specified within the pair ‘<>’ (less than and greater than symbols), the file is searched
only in the standard directories. If the included file is not found, an error is reported and the
compilation stops.

We may create our own header files to be included within our source code files. When we have
functions, structures, type-defined identifiers, macros (defined values) etc., which are needed across
multiple source code files within our project, we send their declarations to header files. Header files act
as common repository for all these declarations. The body of a declared function (definition) may be

241
Quick & Indepth C With Data Structures

written in any one of the source code files of the project. Due to the function declaration being present
in the header file, including the header file allows us to call the function from a different source code file
(other than the one where it is defined). The declaration provides the compiler with the information
regarding the function name, the arguments accepted and its return type. The compiler does not check
for the existence of the body (definition) of the function, which is the responsibility of the linker. The
linker links all the object files (an object file is the compiled representation of a source code file)
together, and checks for the existence of the function implementations in the combined (linked) output.

Assume we have a project with two source code files (source1.c and source2.c) and one header file
(header1.h). We have one function (MyFunction1) which has been defined in ‘source1.c’. Another
function ‘MyFunction2’ is defined in ‘source2.c’ but needs to call the function ‘myFunction1’ to perform
certain tasks. As ‘myFunction1’ is not present in the source file ‘source2.c’, the compiler will raise an
error if we try to call the function from ‘myFunction2’. To overcome the error, we may specify the
declaration of the function ‘myFunction1’ within the header file ‘header1.h’, and include that header file
within the source code file (source2.c) from where the call to the function is made. Similarly, structures,
unions, type-defined values or macros (defined values) may be defined in a header file, which in turn is
included in the source code files wherever they are used.

15.3 MACRO DIRECTIVES

We have already seen the use of ‘#define’ statement which replaces a defined name by a token wherever
the name appears within the source code file. The use of the ‘#define’ statement to define the name
corresponding to the token is called Macro Definition (or simply a macro), and the process of
substituting the defined name with the token within the source code file is called Macro Substitution.
The preprocessor is responsible for achieving the macro substitution by replacing all occurrences of the
defined name with the corresponding token. Let us reiterate the general form of a ‘#define’ statement
as:
#define symbolic-name value

Wherever the symbolic-name appears in the source code file, it is replaced with the value specified
corresponding to the name. The value can be a string, a character, an integer, a real value, a group of
code statements or even another macro. There must be at-least one space each between ‘#define’,
symbolic name and value. The symbolic name must be a valid C name.

The different forms of macro substitution are:

1. Simple macro substitution
2. Macro substitution using arguments
3. Nested macro substitution
4. Standard Predefined Macros

15.3.1 SIMPLE MACRO SUBSTITUTION

We have been using simple macro substitution throughout our book. We associate a simple token (like a
number, character or a string) with a defined name. Wherever the preprocessor finds that name within

242
 PREPROCESSOR DIRECTIVES

the source code file, it replaces that name with the specified token. The modified code is then sent to
the compiler for compilation. Have a look at the next example:
#include <stdio.h>
#define MAX_ELEMENTS 10
int main(void)
{
int rgNumbers[MAX_ELEMENTS] = { 0 };
int idx = 0;
for (; idx < MAX_ELEMENTS; idx++)
{
rgNumbers[idx] = idx * idx;
}
return 0;
}

In the above program, the preprocessor replaces the defined name 'MAX_ELEMENTS' with 10 wherever
it finds that name. So, essentially we have declared an array 'rgNumbers' of 10 elements. Also, the
comparison 'idx < MAX_ELEMENTS' is expanded to 'idx < 10' by the preprocessor before the compiler
takes over.

It is a convention to write the macro name in all capital letters. This helps us to identify them as
symbolic constants and distinguish them from other variables. Though, within the source code, the
macro name is substituted with the expression with which it is associated, a similar text (as the macro
name) appearing within a string are not replaced. If we have a macro as below:
#define MAXNUM 5
All occurrences of ‘MAXNUM’ will be replaced by 5, starting after the line of definition of the macro to
the end of the source code file. If we have code segment as below:
int iTotal = MAXNUM * iRate;
printf(“MAXNUM = %d.”, MAXNUM);

The above code segment will be changed to the following during preprocessing:
int iTotal = 5 * iRate;
printf(“MAXNUM = %d.”, 5);

Notice that the string ‘MAXNUM = %d.’ is left unchanged.

A macro definition may have something more than a simple constant value. It may include expressions
which expand to meaningful expressions wherever they are used. They may include other macro names
too. Some examples are:
#define PI 3.141592
#define DOUBLE_PI (PI * 2)
#define MAXVAL 8 * 64
#define INT_SIZE sizeof(int)

243
Quick & Indepth C With Data Structures

The second macro (DOUBLE_PI) above, uses the previously defined macro ‘PI’ in its expression. So,
wherever we use the macro name ‘DOUBLE_PI’, it is first replaced by the expression ‘(PI * 2)’ and then
to ‘(3.141592 * 2)’. Similarly, wherever we use the macro name ‘INT_SIZE’, it is replaced by the
expression ‘sizeof(int)’.

We may pretty much use any expression within a macro, which replaces the macro name wherever the
macro name is used. Some more interesting macros are:
#define PRINTHELLO printf(“Hello”)
#define GREATER_THAN >
#define EQUAL_TO ==
#define LESS_THAN <
#define INCREMENT ++

The above macros can be used as below:

#include <stdio.h>
int main(void)
{
int iCount = 0;
PRINTHELLO;
if (10 GREATER_THAN 5)
INCREMENT iCount;
if (5 EQUAL_TO 5)
INCREMENT iCount;
}

15.3.2 MACRO SUBSTITUTION USING ARGUMENTS

We can define macros whose use look like a function call. We use the same '#define' directive, but put a
pair of parentheses immediately following the macro name. These type of macros may or may not
accept arguments. To define a macro that uses arguments, we specify parameters between the pair of
parentheses in the macro definition. These parameters must be valid C identifiers, separated by
commas. White space character may optionally be present between the parameters, but, no white space
is allowed between the macro name and the left parenthesis. Consider the below example where the
macro finds the greater number between two numbers.
#define MAX(x, y) (x > y)?x:y
int main(void)
{
printf(“The greater number is %d.”, MAX(5, 7));
}

To invoke a macro that accepts arguments, we write the name of the macro followed by a list of actual
arguments within the parentheses, separated by commas. The number of actual arguments must equal
the number of formal parameters specified in the macro definition. Invoking a macro is also known as a

244
 PREPROCESSOR DIRECTIVES

macro call (similar to a function call). When a macro is called, the preprocessor substitutes the
expression, replacing the formal parameters with the actual parameters. So in our above example, when
we make the macro call ‘MAX(5, 7)’, the macro expression ‘(x > y)?x:y’ is replaced by ‘(5 > 7)?5:7’.

We should be very careful in making a macro call as macro argument replacements may result into a
value which we did not expect. The macro arguments are not evaluated before expanding the macro.
This can be explained using a very simple example:
#define CUBE(x) (x * x * x)
int main(void)
{
printf(“The cube is: %d.\n\n”, CUBE(3 + 2));
return 0;
}

What do we expect when we run the above program? We expect to print the result as 125 (5 * 5 * 5).
But, what we will actually print is 17. This is because the preprocessor performs a blind text
substitution of the formal parameter ‘x’ with the actual parameter ‘3 + 2’. So, what we essentially get
after the text substitution is ‘3 + 2 * 3 + 2 * 3 + 2’, which gives us the final result as 17. This is not the
result we expected! To achieve the desired result, we need to enclose the actual parameters within
parentheses. This changes our macro call to ‘CUBE((3 + 2))’. Notice the extra parenthesis surrounding
the ‘3 + 2’. What we get after the macro substitution in this case is:
‘(3 + 2) * (3 + 2) * (3 + 2)’
This gives us the expected result. But, macro calls may be present in a lot of places within the program.
Missing the parentheses even in a single place will produce erroneous result for our program. An
effective way to make the macro call error free is to change the macro itself as below:
#define CUBE(x) ((x) * (x) * (x))
This makes our macro to be expanded properly, even if we miss the extra parentheses during the macro
call.

The arguments passed to macros can be concatenated using token pasting operator (##). When a
macro is expanded, the two tokens on either side of the ‘##’ operator are combined into a single token.
It is also possible to concatenate two numbers to form a single number. Consider the below example:

#include <stdio.h>
#define MERGE(x, y) x##y
int main()
{
printf("%d\n", MERGE(76, 81));
printf("%s\n\n", MERGE("I Love ", "C Programming"));
}

Output:
7681
I Love C Programming

C provides another operator called stringizing operator (#) which converts macro parameters to string

245
Quick & Indepth C With Data Structures

literals. The actual parameters to the macro are not changed or affected. Wherever the stringizing
operator precedes the use of the argument within the macro, it transforms that to a string literal. All
other locations pertaining to the use of the argument without a preceding stringizing operator remain
unchanged and follow their original behavior.
#include <stdio.h>

#define printline(x) printf(#x "\n")

#define printvalue(x) printf(#x " is %d\n", x)

int main(void)
{
int a = 5;
int b = 10;

printline(I love football);

printvalue(a + b);

return 0;
}

Output:
I love football
a + b is 15

If we have multiple lines of expression in our macro, each line must end with a backslash '\'. The last
line of the expression does not need to have the backslash '\'. Have a look at the next example:

#include <stdio.h>
#include <stdlib.h>

#define FREE_ARRAY(x) if (x != NULL) \

{ \
free(x); \
x = NULL; \
}

int main(void)
{
int i = 0;
int *ptr = (int *)malloc(10 * sizeof(int));
if (ptr)
{
for (; i < 10; i++)
ptr[i] = i;
}

FREE_ARRAY(ptr);
return 0;
}

15.3.3 NESTED MACRO SUBSTITUTION

We may have a macro which is dependent on the outcome of another macro. We may use the other
macro within the definition of this macro. In other words, we may have nested macros wherein the
macro definition of one macro contains the use of other macros. Consider the following examples:

246
 PREPROCESSOR DIRECTIVES

#define NUM_ROWS 10
#define NUM_COLS 4
#define TOTAL_ELEMENTS (NUM_ROWS * NUM_COLS)
#define SQUARE(x) (x * x)
#define CUBE(x) (SQUARE(x) * x)
#define QUAD(x) (CUBE(x) * x)

The macro which is used within another macro needs to be defined first. For the containing macro, the
preprocessor expands each sub (contained) macro until there are no more macros to expand. If we are
using the macro ‘QUAD’ in our code, it will be expanded in the following order:
• The first expansion of ‘(CUBE(x) * x)’ yields the following:
((SQUARE(x) * x) * x), where ‘CUBE(x)’ is expanded to ‘(SQUARE(x) * x)’.
• The second expansion of ‘((SQUARE(x) * x) * x)’ yields the following:
(((x * x) * x) * x), where ‘SQUARE(x)’ is expanded to ‘(x * x)’.

Macros can also be used as parameters of other macros. For example, we may do the following:
#define SQUARE(x) (x * x)
#define HALF(x) (x / 2)
#define HALF_SQUARE(x) HALF(SQUARE(x))

The macro ‘HALF_SQUARE’ finds the half of the square of a number. So, if we provide 6 as the argument
to the macro, it will generate the result as ‘((6 * 6) / 2)’ = 18. This means the output of ‘SQUARE’ will
act as the input of ‘HALF’ producing the result for ‘HALF_SQUARE’.

Similarly, we may nest a macro within itself during the macro call. Few paragraphs ago, we had
discussed a macro ‘MAX’ which finds the maximum of two numbers. We may modify its call to find the
maximum of three numbers as below:
MAX(x, MAX(y, z))

15.3.4 STANDARD PREDEFINED MACROS

There are certain predefined macros which are defined by most C compilers. These macros help achieve
certain tasks easily. Some are:
• __FILE__: This macro expands to the name of the current source file, in the form of a C string
constant. It specifies the name of the file in which this macro is specified (called).
• __FUNCTION__: This macro expands to a C string constant constituting the name of the
function in which this macro is specified.
• __LINE__: This macro expands to the current line number within the source file, in the form of
an integer constant. It specifies the line number of the location where this macro is specified.
• __DATE__: This macro expands to the date on which the preprocessor is being run, in the form
of a C string constant. So, generally this expands to the date of compilation of the source file in
which this macro is specified.
• __TIME__: This macro expands to the time at which the preprocessor is being run, in the form
of a C string constant. So, generally this expands to the time of compilation of the source file in

247
Quick & Indepth C With Data Structures

which this macro is specified.

Usage example:
printf("File: %s, Function: %s, Line: %d, Date and time of compilation: %s %s.\n\n",
__FILE__, __FUNCTION__, __LINE__, __DATE__, __TIME__);

15.3.5 UNDEFINING A MACRO

If a macro ceases to be useful, we may undefine it with the ‘#undef’ directive. We may also require a
macro to behave differently from a different point within the program. For such situations, we may
undefine the macro and redefine it with the new meaning. '#undef' takes a single argument, which is
the name of the macro to undefine. It takes the following form:
#undef MACRO_NAME

15.4 COMPILER CONTROL DIRECTIVES

Compiler control preprocessor directives are statements which change the behavior of the compiler
when compiling the program. The compiler will work according to the instructions specified using the
directives. The compiler control preprocessor directives are of the following types:
• Conditional preprocessor statements
• Pragma directives
• Diagnostic preprocessor statements

15.4.1 CONDITIONAL PREPROCESSOR STATEMENTS

A conditional preprocessor statement is a directive that instructs the preprocessor to include or exclude
a piece of code in the final output code to be passed onto the compiler. A conditional preprocessor
statement is very similar to an 'if' statement in C, except for the fact that an 'if' statement is evaluated
during execution of the program whereas a conditional preprocessor statement is evaluated before a
source code is sent for compilation. The purpose of conditional preprocessor statement is to allow
different code segments to be included in the program depending on certain conditions (like a defined
value etc.) at the time of compilation.

A program may need to use different set of code depending on the machine or operating system in
which it is going to run. The code for one operating system may not work on another operating system
(operating system specific APIs etc.). We use conditional compilation to avoid the portion of the code
which will not work on a specific operating system. The conditional compilation is achieved using a
preprocessor condition where the invalid code can be avoided from the program.

A preprocessor condition whose result is always false is also used to exclude code (code which is not
required right now) from the program but keep it as a sort of comment for future reference.

Now let us discuss the different types of conditional preprocessor directives:

1. #ifdef: This is the simplest form of conditional preprocessor directive. It is paired with a
'#endif' directive, and consists of a block of code (controlled code) specified between the

248
 PREPROCESSOR DIRECTIVES

(#ifdef-#endif) pair. The general form of such a directive is given below:

#ifdef MACRONAME
Controlled code
#endif

The controlled code will be included in the output of the preprocessor (and thus compiled) if
and only if 'MACRONAME' is defined. 'MACRONAME' must be defined in the source file before the
'#ifdef' check, for the conditional check to be successful. The controlled code may include
further preprocessor statements including other conditional preprocessor directives. This
means that just like ‘if’ statements, we may have nested conditional preprocessor statements.
The inner conditional statements are evaluated only if the outer conditions succeed. Just like
the curly braces ('{' and '}') in the nested 'if' statements, '#endif' always matches the nearest
'#ifdef'.

In the below example, the controlled code within (#ifdef-#endif) will get executed as
'MAX_BUFFER' is defined. Commenting or removing the 'MAX_BUFFER' definition will make the
controlled code to be skipped from compilation.
#define MAX_BUFFER 128

#ifdef MAX_BUFFER

#define NUM_ELEMENTS MAX_BUFFER/sizeof(int)

int g_rgiValues[NUM_ELEMENTS];

#endif

2. #ifndef: This is very similar in usage to '#ifdef', except for the fact that it works just the
opposite. In this case, the controlled code is included only if 'MACRONAME' is not defined.
#ifndef MACRONAME
Controlled code
#endif

3. #if: Unlike '#ifdef' which only checks for the existence of a macro, '#if' directive allows us to test
the value of an arithmetic expression. If the value of the expression evaluates to 'true' (non
zero), the controlled code within the '#if-#endif' pair is included in compilation. The general
form of the directive is:
#if expression
Controlled code
#endif

'expression' is a valid C expression, which may contain:

a) Integer constants.
b) Character constants.

249
Quick & Indepth C With Data Structures

c) Arithmetic, bitwise, relational and logical operators.

d) Macros. All macros in the expression are expanded before actual computation of the
expression's value is performed.
e) Use of defined operator, which lets us check whether macros are defined or not.

4. #if-#else: '#else' acts as an extension to the '#if' directive. '#else' also contains a controlled
code, which is sent for compilation only if the '#if' conditional check fails. The general form of
the directive is:
#if expression
Controlled code 1
#else
Controlled code 2
#endif

If 'expression' evaluates to true (non zero), 'Controlled code 1' is compiled, else, 'Controlled
code 2' is compiled.

5. #if-#elif-#else: '#if-#elif-#else' is a further extension of the '#if' directive. '#elif' behaves very
similarly to 'else if' statement, except for the fact that it is a preprocessor directive and the
rules corresponding to '#if' directive must apply to '#elif' too. This means the 'expression'
corresponding to the '#elif' directive must follow the same rules as the '#if' directive. A '#elif'
directive is evaluated only if all previous '#if' and '#elif' directives evaluate to zero (the outcome
of their corresponding expressions have evaluated to false). The general form of the '#elif'
directive is:
#if expression1
Controlled code 1

#elif expression2
Controlled code 2

#elif expression3 /* Optional block */

Controlled code 3

#else
Controlled code 4
#endif

'expression3' is evaluated only if the result of both expressions 'expression1' and 'expression2'
have evaluated to zero. Subsequently, if 'expression3' evaluates to true (non zero), the
'Controlled code 3' gets compiled. 'Controlled code 4' is included in compilation only if all
previous '#if' and '#elif' directives evaluate to zero.

Usage example:

250
 PREPROCESSOR DIRECTIVES

#define LOGLEVEL 7

#if LOGLEVEL == 0
#define LOG_BUFFER_SIZE 10

#elif LOGLEVEL < 5

#define LOG_BUFFER_SIZE 50

#elif LOGLEVEL >= 5 && LOGLEVEL <= 10

#define LOG_BUFFER_SIZE 100

#else
#define LOG_BUFFER_SIZE 200

#endif

6. defined: The 'defined' operator is used in '#if' and '#elif' expressions to test whether a name is
defined as a macro. It evaluates to non zero, if 'MACRONAME' is defined as a macro and zero, if
it is not defined. Thus, '#if defined' MACRONAME is precisely equivalent to '#ifdef' MACRONAME.
Usage example:
#if defined (MACRONAME1) || defined (MACRONAME2)
will evaluate to non zero, if either of the names 'MACRONAME1' or 'MACRONAME2' is defined as
a macro.

15.4.2 Pragma DIRECTIVES

The '#pragma' directive is used to access compiler specific preprocessor extensions. It is used to
specify how a compiler should process its input. The '#pragma' directives are compiler specific and may
vary from one compiler to another. Some common '#pragma' statements are:
1. #pragma once: It indicates that the file, in which the pragma is stated, will be included only
once by the compiler during a build. It is mainly stated at the start of a header file to instruct
the compiler to include that header file just once during the compilation of a source file.

This can be explained using an example. Say, we have two header file 'header1.h' and
'header2.h' and a source file 'source1.c'. The header file 'header2.h' in turn includes 'header1.h'
and the source file includes both the header files. So, the source file gets the reference to
'header1.h' twice, once when it directly includes 'header1.h' and the next via 'header2.h'
(header2.h includes header1.h). This can result into compilation error as the compiler will get
reference to all declarations declared in 'header1.h' twice. One way to prevent the compiler
error is to use this directive at the start of the header file 'header1.h', which makes the
'header1.h' to be included only once during the build.

Another way in which we can achieve this same objective (without the use of #pragma once) is
to put all declarations in 'header1.h' (entire header1.h contents) within a '#ifndef' directive. But,

251
Quick & Indepth C With Data Structures

clearly using the statement ‘#pragma once’ is much simpler than doing the following:
#ifndef _HEADER1_H_INCLUDED_
#define _HEADER1_H_INCLUDED_
/* Specify all declarations here */
#endif

2. #pragma comment: Places a comment into an object or executable file. It has the following
form:
#pragma comment(Comment-Type, Comment-String)
The 'Comment-Type' can be the following:
a) compiler – Places the name and version number of the compiler in the object file.
'Comment-String' is not used and ignored.
b) exestr – Places 'Comment-String' in the executable file. 'Comment-String' is any
programmer specified string. This is mostly used to put copyright information within an
object or executable file.
c) lib – Links the specified library with the executable. The 'Comment-String' parameter
specifies the name (and possibly the path) of the library to link.
d) linker – Places a linker option within the object file. The 'Comment-String' specifies the
linker option.
e) user – Places a general comment in the object file. The 'Comment-String' contains the text
of the comment.

Usage examples:
#pragma comment(compiler)
#pragma comment(lib, "mylib.lib")

3. #pragma warning: Allows us to selectively modify the behavior of compiler warning messages
during compilation. It has the following form:
#pragma warning(Warning-Specifier : Warning-Number-list)
The 'Warning-Specifier' can be the following:
a) once – Makes the compiler display the warning message corresponding to the 'Warning-
Number' only once during the compilation. Example: #pragma warning(once : 4381)
displays the warning 4381 only once.
b) disable – Stops the compiler from displaying the listed warnings. The warning numbers are
specified in the 'Warning-Number-list'. Example: #pragma warning(disable : 4327 154)
prevents the compiler from showing the warnings 4327 and 154.
c) error – Forces the compiler to report the specified warnings as errors. Example: #pragma
warning(error : 4257) makes the compiler report 4257 as an error (instead of a warning).
This forces the compilation to stop immediately.

C99 introduced the _Pragma operator to address a major problem with '#pragma'. '#pragma' being a
directive, cannot be produced as the result of a macro expansion. So, '#pragma' statements cannot be
used within macros. Being an operator, _Pragma can be embedded within macros. It has the following

252
 PREPROCESSOR DIRECTIVES

syntax:
_Pragma(String-Literal)
where 'String-Literal' can be either a normal or wide-character string enclosed within double quotes. It
is a compiler specific directive which may vary from one compiler to another.

15.4.3 DIAGNOSTIC PREPROCESSOR STATEMENTS

Diagnostic preprocessor statements are of two types:
1. #error: The directive '#error' causes the preprocessor to report a fatal error and stop
compilation. The string following '#error' is used as the error message. Example:
#ifdef _32BIT_
#error "This program will not work on 32 bit architecture."
#endif

2. #warning: It is similar to '#error', but causes the preprocessor to issue a warning and continue
the processing.

15.5 SAMPLE PROGRAM

1. N-Queens problem deals with the placing of N chess queens on a N × N Chessboard so that no
two queens can attack each other i.e. no two queens are in the same line diagonally,
horizontally or vertically. We write a program which solves the problem for a defined 'N'. The
chess board is visualized as a two-dimensional N x N array where an element value of '1'
signifies a queen's position. For example following is the array representation for a 4-Queen
solution:
{ 0, 1, 0, 0}
{ 0, 0, 0, 1}
{ 1, 0, 0, 0}
{ 0, 0, 1, 0}

We use the backtracking algorithm for solving the N-Queens problem. We will place queens one
by one in different columns, starting from the left-most column. Before placing a queen in a
column, we will check for clashes (diagonally and horizontally) with already placed queens. We
do not require to check vertically as we place only a single queen in each column eliminating
the possibility of two queens appearing in the same column. In the current column 'n', we try
each row to find a row for which there is no clash.
a) If we find such a row, we mark this array element '[row][column]' with '1' i.e. as part of the
solution and we do a recursive call to find the queen placements for the next remaining
columns.
b) If we do not find such a row for the entire current column 'n', we return false. During our
recursive call 'n - 1' (while processing the previous column 'n - 1'), we get the result from
the nth recursion as false, and we backtrack. We then try to find a different row for the 'n -
1' column (than the one which was previously found) with no clash.

253
Quick & Indepth C With Data Structures

If we do not find any, we return false. In turn, the recursive call for the 'n - 2' column gets
the result as false and that too backtracks and tries to find a different row (than the one
which was previously found). The backtrack continues until a column finds a different row
with no clash and we again start moving forward.

The program prints a 'Q' at the location where a queen is placed and a '#' in all other locations.

#include <stdio.h>

#define NUM_QUEENS 7

/* If your compiler does not support C99 or above, comment the following. */
#define C99_SUPPORTED

#ifndef C99_SUPPORTED
#define _Bool int
#endif

#ifndef false
#define false 0
#endif

#ifndef true
#define true 1
#endif

_Bool CanPlace(unsigned int rgiBoard[NUM_QUEENS][NUM_QUEENS],

int iRow,
int iColumn)
{
/* Variable declaration and initialization. */
int i = 0;
int j = 0;

/* Check if a queen appears in this same row which means a clash. We need to
check only till 'iColumn' as all later columns are yet to be processed. */
for (j = 0; j < iColumn; j++)
{
if (0 != rgiBoard[iRow][j])
{
/* Its a clash. A queen appears in this row but in a different column. */
return false;
}
}

/* Check upper-left diagonal of this element for a clash. */

for (i = iRow, j = iColumn; i >= 0 && j >= 0; i--, j--)
{
if (0 != rgiBoard[i][j])
{
/* There is a clash in the upper-left diagonal. */
return false;
}
}

/* Check lower-left diagonal of this element for a clash. */

for (i = iRow, j = iColumn; i < NUM_QUEENS && j >= 0; i++, j--)
{
if (0 != rgiBoard[i][j])
{

254
 PREPROCESSOR DIRECTIVES

/* There is a clash in the lower-left diagonal. */

return false;
}
}

/* There are no clash for this row. */

return true;
}

_Bool PlaceQueen(unsigned int rgiBoard[NUM_QUEENS][NUM_QUEENS],

int iColumn)
{
/* Variable declaration and initialization. */
int iRow = 0;
_Bool fPlaced = false;

/* All queens are already placed. Return true. */

if (iColumn >= NUM_QUEENS)
{
return true;
}

/* Try to find a row for this column where we will have no clash. */
for (; iRow < NUM_QUEENS; iRow++)
{
/* Check if the queen can be placed in this row. */
if (CanPlace(rgiBoard, iRow, iColumn))
{
/* We found a row with no clash. Mark this array element as a
possible solution. Note: This is not the final solution as it
may change if no possible location is found in a later column. */
rgiBoard[iRow][iColumn] = 1;

/* Use recursion to find the queen placements for the

remaining columns to the right. */
if (PlaceQueen(rgiBoard, iColumn + 1))
{
/* This column and all later columns are solved. So, this row is
the correct position for this column and we abondon the search. */
fPlaced = true;
break; /* break out of the for loop. */
}

/* Queen's placement could not be found in one of the later columns. So

in this column, we need to find a different row for the queen which
may resolve the clash in the later column. In a single word: BACKTRACK */
rgiBoard[iRow][iColumn] = 0;
}
}

return fPlaced;
}

void PrintBoard(unsigned int rgiBoard[NUM_QUEENS][NUM_QUEENS])

{
/* Variable declaration and initialization. */
int iRow = 0;
int iColumn = 0;

printf("The solution for the %u-Queens problem is:", NUM_QUEENS);

for (; iRow < NUM_QUEENS; iRow++)
{
printf("\n");

255
Quick & Indepth C With Data Structures

for (iColumn = 0; iColumn < NUM_QUEENS; iColumn++)

{
printf("%c ", (0 != rgiBoard[iRow][iColumn]) ? 'Q' : '#');
}
}
}

int main(void)
{
/* Variable declaration and initialization. */
unsigned int rgiBoard[NUM_QUEENS][NUM_QUEENS] = { 0 };

if (!PlaceQueen(rgiBoard, 0))
{
/* Returned 0 (false). ['!' of '!= 0' means '== 0'] */
printf("Failed to find a solution to the %u-Queens problem.\n\n",
NUM_QUEENS);
}
else
{
/* We found a solution. Print the result. */
PrintBoard(rgiBoard);
printf("\n\n");
}

return 0;
}

256
 DEPRECATED FUNCTIONS

DEPRECATED FUNCTIONS

[CHAPTER-16]
Just like most other computer languages, advancements have been made in C too, with newer versions
released over the years. Each new version has come with new features and most importantly security
enhancements. Many old library functions had security flaws which made them prone to memory faults,
hack attacks and overflow issues. Simple programming error could lead these functions to go kaput
leading to the program crash. As these functions were already in use by many applications around the
world, changing their behavior, their signature (prototype) or removing them entirely were not feasible.
So, newer version of these functions were developed where the functions had a similar but different
name and were recommended to be used in new applications. Though the signature was changed for
many of the new version functions, the changes were very limited and bore resemblance to the older
ones. The old functions were deprecated and not recommended to be used any more. Though we may
still use these functions, but, the compiler will generate a warning if we try to do so. We must remember
that future versions of C may remove these functions entirely. Some of the library functions which have
been deprecated and their newer secure counterparts are given below:

Sr. No. Deprecated Recommended Deprecated Recommended

(ANSI Version) (ANSI Version) (Wide Char Version) (Wide Char Version)
1 asctime asctime_s _wasctime _wasctime_s
2 atof strtod _wtof wcstod
3 atoi strtol _wtoi wcstol
4 atol strtol _wtol wcstol
5 atoll strtoll _wtoll wcstoll
6 bsearch bsearch_s
7 ctime ctime_s _wctime _wctime_s
8 fopen fopen_s _wfopen _wfopen_s
9 fprintf fprintf_s fwprintf fwprintf_s
10 freopen freopen_s _wfreopen _wfreopen_s
11 fscanf fscanf_s fwscanf fwscanf_s
12 getenv getenv_s _wgetenv _wgetenv_s
13 gets gets_s _getws _getws_s
14 gmtime gmtime_s
15 localtime localtime_s
16 mbsrtowcs mbsrtowcs_s
17 mbstowcs mbstowcs_s
18 memcpy memcpy_s wmemcpy wmemcpy_s
19 memmove memmove_s wmemmove wmemmove_s
20 printf printf_s wprintf wprintf_s
21 qsort qsort_s

257
Quick & Indepth C With Data Structures

Sr. No. Deprecated Recommended Deprecated Recommended

(ANSI Version) (ANSI Version) (Wide Char Version) (Wide Char Version)
22 setbuf setvbuf
23 snprintf _snprintf_s _snwprintf _snwprintf_s
24 sprintf sprintf_s swprintf swprintf_s
25 sscanf sscanf_s swscanf swscanf_s
26 strcat strcat_s wcscat wcscat_s
27 strcpy strcpy_s wcscpy wcscpy_s
28 strerror strerror_s _wcserror _wcserror_s
29 strncat strncat_s wcsncat wcsncat_s
30 strncpy strncpy_s wcsncpy wcsncpy_s
31 strtok strtok_s wcstok wcstok_s
32 vfprintf vfprintf_s vfwprintf vfwprintf_s
33 vfscanf vfscanf_s vfwscanf vfwscanf_s
34 vprintf vprintf_s vwprintf vwprintf_s
35 vscanf vscanf_s vwscanf vwscanf_s
36 vsnprintf vsnprintf_s _vsnwprintf _vsnwprintf_s
37 vsprintf vsprintf_s vswprintf vswprintf_s
38 vsscanf vsscanf_s vswscanf vswscanf_s

258
 BEST PROGRAMMING PRACTICES

BEST PROGRAMMING PRACTICES

[CHAPTER-17]

17.1 INTRODUCTION
As programmers we should always try to write code in a way that makes it:
• Better readable
• Easily understandable
• Better manageable
• Easily maintainable
• Prevent errors from creeping in
• Make our programs more robust

In this chapter, we will understand the techniques which help achieve the above points. Lets start…

17.1.1 DOCUMENTATION
Provide adequate documentation within the entire code using the help of comments. At every critical
code segment, we should explain the reason for doing something in a particular way. Comments should
be provided explaining the logic and the control flow. Most programs require future maintenance where
they need to be reworked (for adding features and fixing bugs). The maintenance may not be done by
the original programmers. The rework may also be performed after a considerable amount of time for
which even the original programmers too lose track and forget the logic used. So, it becomes essential
that we make the program understandable for future reference. But, we should not become over
enthusiastic and fill our program with more comments than the actual code :-) itself. Documentation
must be just enough to make the program better understandable. Too much documentation will reduce
the readability of the program and too less documentation will make it difficult to be understood.
Documentation helps make the program understandable, manageable and maintainable.

17.1.2 AVOID USING DEPRECATED FUNCTIONS

Use of deprecated functions make our program error prone and susceptible to hacker and virus attacks.
So, always try to use the newer version of the functions. See the chapter on ‘Deprecated Functions’ for
a list of deprecated functions and their recommended alternatives. This technique prevents errors from
creeping in our program and makes it more robust.

17.1.3 TABS
The source code editor (IDE) should be configured to replace tabs with spaces. Tab stops should be set
at four spaces. This ensures that we will see the code in the same way irrespective of the editor in which
they are viewed.

259
Quick & Indepth C With Data Structures

17.1.4 INDENTATION
Every new scope should be indented by four spaces (one tab stop). Every time a new scope is
introduced, either through the use of a compound statement or the use of braces, indentation of all the
statements contained within the new scope should be increased by one (four spaces).

17.1.5 LINE LENGTH

No line should exceed the width of a typical source code editor which is approximately 80-100
characters. A longer line should be broken into multiple lines using line breaks.

17.1.6 BLANK LINES

We can use a single blank line to separate areas of functionality within a method. For example, we could
use a blank line after a loop construct to separate it from the below statements. Two (or three) blank
lines can be used to separate the end of one method from the beginning of another.

17.1.7 FILE HEADER

Every file (whether header or source) should start with a comment header. The header should explain
the purpose of the file, its creation date, modification history and copyright information (if any). The
format of the header must be consistent across all the files. A file header may look as below:
/*--------------------------------------------------------------------------------
** File: MySource.c
**
** Desc: Description of the file. Long description should be broken into multiple
** lines and also indented like this.
**
** 2017/06/19 - Sudripta Nandy - Initial implementation.
** 2017/11/17 – Sudripta Nandy – Fixed bug #68.
**
** Copyright (c) 2017-2018, MyCompany. All rights reserved.
**--------------------------------------------------------------------------------*/

17.1.8 FUNCTION HEADER

Every function should start with a comment header. The header should explain the purpose of the
function, the description of the parameters it accepts, and the possible return values.
/*--------------------------------------------------------------------------------
** Name: MyFunction
** Desc: The description of the function. Long description should be broken into
** multiple lines and also indented like this.
**
** Parameters:
** Parameter1: Description of first parameter.
** Parameter2: Description of second parameter. Long description should
** be broken into multiple lines and also indented like this.
**
** Returns:
** 0: Description of when it returns 0.
** 1: Description of when it returns 1.
//--------------------------------------------------------------------------------*/

260
 BEST PROGRAMMING PRACTICES

17.1.9 BRACES
When a code block starts, we may opt to place the opening brace in the same line as the start of the
block statement (if, for, while etc.) or as separately on the following line. The former option is called the
K&R style, and the latter is called the Allman style. Either style is acceptable, but it should be used
consistently across the entire program. The K&R style makes the code more compact but at the cost of
readability as the opening brace may be easily overlooked. The Allman style makes the opening and
closing braces clearly visible.
/* K&R Style */
if (x < 5) {
printf("x is less than 5");
}

/* Allman Style */
while (x > 0)
{
x--;
}

17.1.10 PREVENT MULTIPLE INCLUSIONS

All include (header) files should be protected from being included multiple times during compilation. It
can be done in various ways. The preferred way is to use the '#pragma once' directive. Please refer to
the topic on '#pragma once' under section 15.4.2.

17.1.11 FUNCTION POINTERS

It is recommended to type-define function pointers. It makes the code cleaner, better readable and
easier to understand.

17.1.12 USE OF PARENTHESIS

Wherever we have doubt about operator precedence or associativity, we may use parenthesis to
sequence the execution of the operations. Use of parenthesis within complex statements also improve
readability allowing us to easily understand the sequence of execution of the operations. When we are
working with bitwise operators, use of parenthesis is highly recommended.
a = ((iValue & 7) << 4);

17.1.13 BOOLEAN COMPARISONS

We should never compare a boolean value for equality or inequality with ‘true’ or 1. Comparisons
should always be done for equality or inequality with ‘false’ or 0 (zero). We must remember that any
non-zero value represents ‘true’. The exact value of ‘true’ may differ between programming languages.
A function which returns a boolean value may return any non-zero value to denote a ‘true’ outcome
(value). So, comparing the returned value with ‘true’ (in the caller function) may give us incorrect
results. Comparisons for equality or inequality with 'false' are permissible because 'false' is always 0.

261
Quick & Indepth C With Data Structures

We may also use the NOT (!) operator for checking equality or inequality with ‘false’.

17.1.14 COMPARING CONSTANTS

When a constant value (or a defined value) is used within a comparison, it should be placed on the left
side of the comparison. This will force the compiler to issue an error when we accidentally use '='
instead of '=='.
if (0 == x) /* Recommended */
if (x == 0) /* Not recommended as we may mistakenly write (x = 0) */

17.1.15 ALWAYS CHECK RETURN VALUES

When a caller calls another function, it must always check the value returned by the called function
before continuing with its execution. The called function might have failed which may result into errors if
further processing continues. In worst cases, the program may hit a memory fault and crash. For this
reason, every function that has a return code should have its return code checked. Though we may have
situations where it is safe to continue even if a called function fails. For rest of the functions, we must
check their return code before continuing.

17.1.16 BLOCK STATEMENTS

All block control statements like conditional or loop constructs (if statements, while statements, for
statements etc.) should be enclosed within opening and closing brace pair. Even if the constructs
contain a single statement, they must be enclosed within braces.
/* Recommended */
if (x < 5)
{
x++;
}
/* Not Recommended */
if (x < 5)
x++;

17.1.17 SWITCH STATEMENTS

A switch statement must be indented as below:
switch (variable)
{
case value1:
….
break;

case value2:
….
break;

262
 BEST PROGRAMMING PRACTICES

default:
….
break;
}

A common bug with 'switch' statements is to forget placing a 'break' after executing a 'case' label. Each
case code block should be ended with a break statement. If our logic desires an intentional fall thru to
the next 'case' block, we must clearly mark the fall thru using a comment.

17.1.18 DISABLING OUT CODE

When a portion of code is not needed right now, but may make a possible comeback in the future,
should not be removed and retained as a non working code. To make a piece of code non working, we
may comment that portion of the code or put that code within a ‘#if’ block (preferable) as below:
#if 0
Obsolete code
#endif

We should, using a comment put an explanation behind the reason for disabling the code.

17.1.19 SHORT FUNCTIONS ARE BEAUTIFUL

Long functions are not only harder to maintain but also difficult to understand. Ideally, functions should
be able to fully fit onto the screen at one time. Functions longer than approximately 50 lines should be
examined to see if they can be broken up into smaller functions. This isn’t a hard and fast rule, but will
produce more maintainable code.

17.1.20 LIMIT EXIT POINTS FROM FUNCTIONS

Having more than one exit point in a function which allocates memory, risks memory leaks. A common
bug is to forget freeing (deallocate) an allocated memory when we return early from a function. When
our function allocates memory from heap using 'malloc' or 'calloc', we will leak memory if we forget
freeing them using 'free'. If we return early from multiple places, we have to remember freeing the
allocations before every point of return. Even if all of these cases are handled, maintenance of the code
may be cumbersome and difficult. So, we should try to limit the number of exit points from a function.

There are multiple ways to limit the number of exit points from a function. We will discuss two of them:
• Using goto: We may put a label before a single exit point of the function. The label will be
followed by all the clean up code (deallocation etc.) required for the function. Every location
which desires a return from the function, will use a goto statement to catapult the control to the
label. The construct will be something similar to the following:
void MyFunc(void)
{
int *ptr = NULL;

263
Quick & Indepth C With Data Structures

ptr = (int *) malloc(10 * sizeof(int));

if (NULL == ptr)
{
goto CLEANUP;
}
….
….
if ( … )
{
goto CLEANUP;
}
….
….
CLEANUP:
if (NULL != ptr)
{
free(ptr);
ptr = NULL;
}
}

• Using do-while construct: We have been using this construct in most of our programs. In this
approach, we put most of our function code (except the variable declarations) within a do-while
construct. This do-while construct is a bit unique in the sense that it does not loop, and its body
is executed just once. We make this a single execution loop by making the loop condition such
that it never fulfills. But being an exit-controlled loop, the loop body gets executed once even
when the loop condition fails. The construct will be something similar to the following:
void MyFunc(void)
{
int *ptr = NULL;
do
{
ptr = (int *) malloc(10 * sizeof(int));
if (NULL == ptr)
{
break;
}
….
….
if ( … )
{
break;
}

264
 BEST PROGRAMMING PRACTICES

….
….
} while(0);

if (NULL != ptr)
{
free(ptr);
ptr = NULL;
}
}

Though both the above approaches work, the second approach is desirable. This is because, the
use of goto brings its own side effects (refer section 6.8). Also, use of goto statements reduces
program optimization leading to inefficient programs.

17.1.21 INITIALIZE VARIABLES

This is one of the most basic rules of C programming. Uninitialized C variables (except static and global
variables) contain garbage values. If we do not initialize them, we may end up comparing them with
incorrect values. Pointer variables must be initialized to NULL. Initializing variables also help us during
clean up of the variables before we exit the function. If a variable is supposed to be allocated, we must
compare its value with NULL before attempting to use it or deallocate it. This prevents us from trying to
access an invalid memory location. After a successful deallocation, we must again make the pointer
variable point to NULL to prevent its further use.

17.1.22 VARIABLE NAMING CONVENTIONS

We should always name variables according to their usage. By looking at a variable name we should be
able to understand its purpose, its scope and its type. We should also be able to understand whether a
variable is a constant, an array, a string or a pointer. This ensures easy understanding of the program.
Some of the variable naming conventions are:
• Always precede global variable names with with a ‘g_’ and static variable names with a ‘s_’.
• A single character variable should start with a ‘c’, an integer variable should start with an ‘i’, a
float variable should start with a ‘flt’, a double should start with a ‘dbl’ and a boolean variable
should start with an ‘f’.
• An array should start with a ‘rg’ and a pointer variable should start with a ‘p’.
• ANSI character string should start with ‘sz’ and a wide character string should start with ‘wsz’.
• Constants should have all UPPERCASE names.
• Most importantly, the name of a variable should indicate its purpose or use.

Examples:
g_fltRate (global float variable), g_rgiRates[10] (global integer array), s_fToggle (static boolean
variable), szName[100] (character string), pszName (pointer to a string), piQuantity (pointer to an
integer), MAX_LENGTH (defined constant value).

265
Quick & Indepth C With Data Structures

LINKED LISTS

[CHAPTER-18]

18.1 INTRODUCTION
A list is a group of items organized sequentially. Arrays are one type of lists. We can access array
elements using their index or by incrementing a pointer to point to the desired element. As the array
elements are sequentially arranged, accessing array elements are quick. But, arrays have some major
disadvantages. We need to know the required size of an array beforehand. An array cannot be resized
as per our growing need. We may resize an array only if we dynamically allocate the array space from
heap using the functions ‘malloc’ or ‘calloc’. But, to resize such an array we need to reallocate the array
to a new location which is not only time consuming but also a processor intensive task. Moreover, there
is no guarantee that we will get enough contiguous free memory space for the new allocation. To
overcome this, if we over predict our requirement when creating the array, we will be wasting memory
and may quickly run out of memory for other tasks.

In many practical applications, it is impossible to predict our requirement at the beginning. We need to
keep allocating as per our requirement. So, arrays make a very inefficient solution to such problems. An
efficient solution can be reached with a very different approach using a new kind of list. In this
approach, we make each item of the list a part of a structure along with a pointer pointing to the next
such structure. In the same sequence, the next structure too contains a link (pointer) to the structure
coming after it. These links allow us to move forward within the list. Optionally, a structure may also
contain another link pointing back to the previous structure (the structure coming before it). This
optional back link allows us to move back within the list.

The above figure shows a typical representation of a single linear linked list. A single data structure
within the list is called a ‘node’. The above figure contains four nodes. A linked list is stated as ‘single’, if
we only have a forward link in the linked list. This means that a node only points to the next node within
the list. A node lacks a pointer to the previous node within the list. A linked list is called as ‘linear’, if the
last node within the list points to NULL, which marks the end of the list.

The first logical node within a list is called the head node and the last logical node within the list is
called the tail node. In the above figure we can see that the head node has a starting memory address
of 128 and the tail node has a starting memory address of 864. Similarly, the two middle nodes have
addresses of 288 and 656 respectively. We can see that the nodes are not allocated consecutive

266
 LINKED LISTS

memory locations. A node is assigned a memory address wherever enough consecutive free space is
available to just accommodate that node itself. The previous node in the list is then made to point to
this newly allocated node.

There are various advantages and disadvantages of using linked lists and they are:

Advantages:
• With arrays we need to have a clear idea regarding the number of elements that are going to be
required. Except for dynamic allocation, arrays cannot be expanded or contracted as per
requirement. Even for dynamically allocated arrays, expanding or contracting the array is a time
consuming and processor intensive task. Linked lists can be easily expanded and contracted as
per requirement. Adding a new node or deleting an existing node are low resource operations
for linked lists.
• Insertion or deletion of a node is much simpler to do in linked lists.
• Array elements are assigned contiguous memory locations for which a large chunk of free
contiguous memory space is required for the allocation to be successful. On the other hand, the
nodes of a linked list are non contiguous in memory location. Each node is allocated a different
location wherever enough free memory capable of containing that node is available. This makes
multiple small memory locations to be allocated for a linked list, unlike a large chunk of
contiguous memory for arrays. So, memory availability is better guaranteed for linked lists.

Disadvantages:
• Unlike arrays, random access of linked list nodes is not possible. We have to access nodes
sequentially starting from the first or last node. For searching and access of elements, arrays
are faster than linked lists.
• As random access is not possible, we cannot do binary search with linked lists.
• Extra memory space for pointers is required with each node of the list.
• Pointer arithmetic for element access can be used with arrays but not with linked list. This again
makes element access and retrieval quicker for arrays.

For situations where the number of elements are fixed with no requirement of insertion or deletion of
elements, arrays are preferred. For all other cases, linked list make a much better choice.

18.2 LINKED LIST TYPES

There are four possible types of linked list, and they are:
1. Single linear linked list
2. Double linear linked list
3. Single circular linked list
4. Double circular linked list

18.2.1 SINGLE LINEAR LINKED LIST

The example linked list discussed in section 18.1 is a single linear linked list. The nodes in this list

267
Quick & Indepth C With Data Structures

contain a single pointer pointing to the next node in the list. These nodes do not have a back pointer
which points to the previous node. This makes the list inoperable for backward traversal, and we can
only move forward within the list. An example of the node structure of a single linear linked list is:
struct ITEMDETAILS
{
int iItemID;
char szItemName[50];
float fltItemRate;
struct ITEMDETAILS *pNext;
};

‘iItemID’, ‘szItemName’ and ‘fltItemRate’ mark the data portion of the node and ‘pNext’ is the pointer to
the next node. Optionally, we may type-define the structure for easy usage across our program. Our
example program (provided a little later), does exactly the same.

Our program will have a pointer (possibly global) which will point to the starting node (Head) of the list.
When the list does not have any nodes, this pointer will point to NULL. Otherwise, this pointer will
always point to the head node of the list. Optionally, we may have a pointer which points to the last
node (Tail) of the list. This pointer too points to NULL when there are no nodes in the list.

Adding nodes to the end of the list is a simple task. If our list is empty, we mark the new node as ‘Head’
node of the list. In case of a non-empty list, we make the ‘pNext’ pointer of the ‘Tail’ node point to the
newly created node. We make the ‘pNext’ pointer of the new node point to NULL and mark the new node
as the new ‘Tail’ node of the list.

Inserting a new node between two existing nodes is a little more tricky task. If a new node needs to be
inserted between two nodes (‘A’ and ‘B’), we make the ‘pNext’ pointer of the previous node (‘A’) point to
the new node and the ‘pNext’ pointer of the new node point to the next node (‘B’).

Deleting an existing node requires us to readjust the ‘pNext’ pointer of the previous node. If the node to

268
 LINKED LISTS

be deleted is the ‘Head’ node of our list, we make the node following the current ‘Head’ node (the node
to be deleted) as the new ‘Head’ node of the list. In all other cases, we make the ‘pNext’ pointer of the
previous node point to the node which comes after the node to be deleted.

Now, let us understand single linear linked list using a program. In the program, we will hold the
electricity consumption details corresponding to the customers of an electric company. The user will
provide the total units consumed in a month by a consumer. The consumers will be identified by a
unique consumer ID. We will allow the user to add/insert node to the list, delete node from the list and
traverse the list.

#include <stdio.h>
#include <stdlib.h>

typedef struct _CONSUMPTIONINFO

{
long lConsumerID;
int iUnitsConsumed;

struct _CONSUMPTIONINFO *pNext;

}CONSUMPTIONINFO, *PCONSUMPTIONINFO;

PCONSUMPTIONINFO pHead = NULL;

/* The function creates a new node and adds it to the list. The function adds the nodes in
ascending order based on the consumer ID. */
void AddNodeSorted(void)
{
CONSUMPTIONINFO *pNewNode = NULL;
CONSUMPTIONINFO *pMove = NULL;

do
{
/* Allocate the new node. */
pNewNode = (CONSUMPTIONINFO *) malloc(sizeof(CONSUMPTIONINFO));
if (NULL == pNewNode)
{
printf("Failed to allocate memory for the new node.\n");
break;
}

/* Accept the consumer details from the user. */

printf("Enter the consumer ID: ");
scanf("%ld", &pNewNode->lConsumerID);

printf("Enter the total units consumed: ");

scanf("%d", &pNewNode->iUnitsConsumed);

if (NULL == pHead || pNewNode->lConsumerID < pHead->lConsumerID)

{
/* If our list is empty, or the consumer ID of the new node is less than the
previous 'Head' node (i.e. the smallest in the list), our new node needs

269
Quick & Indepth C With Data Structures

to be the new 'Head' node of the list.

If our list was empty, the 'pNext' pointer of the new node must point to
NULL. This is because the newly created node is not only the first node,
but also the last node of the list. In case of an empty list 'pHead' is
NULL, which makes the statement 'pNewNode->pNext = pHead' the same as
'pNewNode->pNext = NULL'. New (Empty list): |N| --> NULL

If the list was non empty and the new node needs to be the new 'Head' node,
the statement 'pNewNode->pNext = pHead' makes the 'pNext' pointer of the
new node point to the previous starting (Head) node. We then make our
'pHead' pointer point to the newly created node, i.e. the newly created
node now becomes the new 'Head' node of the list.
Previously: |3| --> |5| --> |9| --> NULL
Updated (New value 2): |2| --> |3| --> |5| --> |9| --> NULL */

pNewNode->pNext = pHead;
pHead = pNewNode;
}
else
{
/* Locate the position to insert the new node. */
for (pMove = pHead; pMove->pNext != NULL; pMove = pMove->pNext)
{
if (pNewNode->lConsumerID < pMove->pNext->lConsumerID)
{
break;
}
}

/* The new node will be placed after 'pMove'. Make the new node point to the
node which 'pMove' was pointing to. We then make 'pMove' point to the new
node. If 'pMove' is at |5|,
Previously: |3| --> |5| --> |9| --> NULL
Updated (New value 7): |3| --> |5| --> |7| --> |9| --> NULL */
pNewNode->pNext = pMove->pNext;
pMove->pNext = pNewNode;
}

printf("Node successfully added\n");

} while (0);
}

/* The function accepts the consumer ID and deletes the first occurring node with that
consumer ID. */
void DeleteNode(void)
{
CONSUMPTIONINFO *pMove = NULL;
CONSUMPTIONINFO *pPrev = NULL;
long lConID = -1;

do
{
if (NULL == pHead)
{
break;
}

/* Accept the consumer ID to delete. */

printf("Enter the consumer ID to delete: ");
scanf("%ld", &lConID);

270
 LINKED LISTS

if (pHead->lConsumerID == lConID)
{
/* We need to delete the 'Head' node. Make the node following the 'Head'
node as the new 'Head' node. If there are no nodes after the 'Head'
node, then 'pHead' will now point to NULL. After update, 'pMove' points
to |3| and 'pHead' points to |5|.
Previously: |3| --> |5| --> |9| --> NULL
Updated (Delete value 3): |5| --> |9| --> NULL */

pMove = pHead;
pHead = pHead->pNext;
}
else
{
/* Locate the node with the consumer ID. 'pPrev' points to the node which
precedes the node to delete. 'pMove' points to the node to delete. */
for (pPrev = pHead, pMove = pHead->pNext; pMove != NULL;
pPrev = pMove, pMove = pMove->pNext)
{
if (pMove->lConsumerID == lConID)
{
break;
}
}

if (NULL != pMove)
{
/* We found the node to delete. Make the previous (preceding)
node (pPrev) point to the node following the node to delete.
'pPrev' points to |3| and 'pMove' points to |5|.
Previously: |3| --> |5| --> |9| --> NULL
Updated (Delete value 5): |3| --> |9| --> NULL */
pPrev->pNext = pMove->pNext;
}
}

if (NULL != pMove)
{
/* We found the node to delete. Free the node memory. */
free(pMove);
pMove = NULL;
printf("Node successfully deleted.\n");
}
else
{
printf("No node with the given consumer ID found.\n");
}

} while (0);
}

/* The function traverses the entire list and prints each node's values. */
void TraverseList(void)
{
CONSUMPTIONINFO *pMove = pHead;

for (; NULL != pMove; pMove = pMove->pNext)

{
printf("Consumer ID: %ld, Units Consumed: %d\n",
pMove->lConsumerID, pMove->iUnitsConsumed);
}
}

271
Quick & Indepth C With Data Structures

int main(void)
{
int iOption = 0;

while (4 != iOption)
{
printf("\n1. Add node\n2. Delete node\n3. Traverse list\n4. Exit");
printf("\n\nEnter your option (1-4): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
AddNodeSorted();
break;
case 2:
DeleteNode();
break;
case 3:
TraverseList();
break;
default:
break;
}
}

/* Memory cleanup before exit. */

for (PCONSUMPTIONINFO pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
}

return 0;
}

18.2.2 DOUBLE LINEAR LINKED LIST

Every node in this list contain two pointers, one pointing to the previous node in the list and another
pointing to the next node. The ‘pPrev’ pointer of the ‘First/Head’ node, and the ‘pNext’ pointer of the
‘Last/Tail’ node point to NULL. As we have a back pointer too in every node within the list, the list
becomes suitable for backward traversal, and we not only can move forward but also backward within
the list. An example of the node structure of a double linear linked list is:
struct ITEMDETAILS
{
int iItemID;
char szItemName[50];
float fltItemRate;
struct ITEMDETAILS *pPrev;
struct ITEMDETAILS *pNext;
};

‘pPrev’ is the pointer to the previous node and ‘pNext’ is the pointer to the next node.

272
 LINKED LISTS

We will have an almost equivalent code as of a single linear linked list when adding a node at the end of
the list. Additionally, we will also make the ‘pPrev’ pointer of our new node point to the previous ‘Tail’
node.

If a new node needs to be inserted between two nodes (‘A’ and ‘B’), we make the ‘pNext’ pointer of the
previous node (‘A’) point to the new node, the ‘pPrev’ pointer of the new node point to the previous
node (‘A’), the ‘pNext’ pointer of the new node point to the next node (‘B’) and the ‘pPrev’ pointer of the
next node (‘B’) point to the new node.

In case of delete, if the node to be deleted is the ‘Head’ node of our list, we make the node following the
current ‘Head’ node as the new ‘Head’ node. If we are deleting a node say ‘B’ situated between nodes
‘A’ and ‘C’, we will make the ‘pNext’ pointer of ‘A’ point to ‘C’, and the ‘pPrev’ pointer of ‘C’ point to ‘A’.
This will make ‘B’ get disconnected from the linked list.

In our next program, we will use the same case situation as the single linear linked list program. The
only difference is that we will now use a double linear linked list instead. The 'TraverseList' function will
be exactly similar to the previous program, and so we are not reproducing the same over here.

273
Quick & Indepth C With Data Structures

#include <stdio.h>
#include <stdlib.h>

typedef struct _CONSUMPTIONINFO

{
long lConsumerID;
int iUnitsConsumed;

struct _CONSUMPTIONINFO *pPrev;

struct _CONSUMPTIONINFO *pNext;

}CONSUMPTIONINFO, *PCONSUMPTIONINFO;

PCONSUMPTIONINFO pHead = NULL;

/* The function creates a new node and adds it to the list. The function adds the nodes
in ascending order based on the consumer ID. */
void AddNodeSorted(void)
{
CONSUMPTIONINFO *pNewNode = NULL;
CONSUMPTIONINFO *pPrevNode = NULL;
CONSUMPTIONINFO *pMove = NULL;

do
{
/* Allocate the new node. */
pNewNode = (CONSUMPTIONINFO *) malloc(sizeof(CONSUMPTIONINFO));
if (NULL == pNewNode)
{
printf("Failed to allocate memory for the new node.\n");
break;
}

/* Accept the consumer details from the user. */

printf("Enter the consumer ID: ");
scanf("%ld", &pNewNode->lConsumerID);

printf("Enter the total units consumed: ");

scanf("%d", &pNewNode->iUnitsConsumed);

if (NULL == pHead || pNewNode->lConsumerID < pHead->lConsumerID)

{
/* If our list is empty, or the consumer ID of the new node is less
than the previous 'Head' node (i.e. the smallest in the list),
our new node needs to be the new 'Head' node of the list. */

/* Make 'pNext' pointer of 'pNewNode' point to the old 'Head' node. */

pNewNode->pNext = pHead;
if (NULL != pHead)
{
/* Make 'pPrev' pointer of the old 'Head' node point to the new node.
This makes our new node get positioned at the start of the list. */
pHead->pPrev = pNewNode;
}

/* As 'pNewNode' now becomes the new starting node,

its 'pPrev' pointer needs to point to NULL. */
pNewNode->pPrev = NULL;
pHead = pNewNode; /* 'pNewNode' is now the new 'Head'. */
}
else

274
 LINKED LISTS

{
/* Locate the position to insert the new node.
The new node will come in between 'pPrevNode' and 'pMove'.
|pPrevNode| --> |pNewNode| --> |pMove| ->
For easier explanation, we are using two pointers 'pPrevNode'
and 'pMove'. This can instead be done using a single pointer. */
for (pPrevNode = pHead, pMove = pHead->pNext; pMove != NULL;
pPrevNode = pMove, pMove = pMove->pNext)
{
if (pNewNode->lConsumerID < pMove->lConsumerID)
{
break;
}
}

/* Make the 'pNext' pointer of the new node point to 'pMove' i.e.
the node (or NULL) which will come after this new node. */
pNewNode->pNext = pMove;
/* Make the 'pPrev' pointer of the new node point to 'pPrevNode'
i.e. the node which will come before this new node. */
pNewNode->pPrev = pPrevNode;
/* As the new node will come after 'pPrevNode', make the 'pNext'
pointer of the previous node 'pPrevNode' point to the new node. */
pPrevNode->pNext = pNewNode;
if (NULL != pMove)
{
/* As the new node will come before 'pMove', make the 'pPrev'
pointer of the next node 'pMove' point to the new node. */
pMove->pPrev = pNewNode;
}
}

printf("Node successfully added\n");

} while (0);
}

/* The function accepts the consumer ID and deletes the first

occurring node with that consumer ID. */
void DeleteNode(void)
{
CONSUMPTIONINFO *pMove = NULL;
CONSUMPTIONINFO *pPrevNode = NULL;
long lConID = -1;

do
{
if (NULL == pHead)
{
break;
}

/* Accept the consumer ID to delete. */

printf("Enter the consumer ID to delete: ");
scanf("%ld", &lConID);

if (pHead->lConsumerID == lConID)
{
/* We need to delete the 'Head' node. Make the node following the
'Head' node as the new 'Head' node. If there are no nodes after
the 'Head' node, then 'pHead' will now point to NULL. */

/* We make 'pMove' point to the node to delete. */

275
Quick & Indepth C With Data Structures

pMove = pHead;
/* The second node becomes the new 'Head' node. */
pHead = pHead->pNext;

if (NULL != pHead)
{
/* Being the new 'Head' node, it's 'pPrev' must now point to NULL. */
pHead->pPrev = NULL;
}
}
else
{
/* Locate the node with the consumer ID.
'pPrevNode' points to the node which precedes the node to delete.
'pMove' points to the node to delete. */
for (pPrevNode = pHead, pMove = pHead->pNext; pMove != NULL;
pPrevNode = pMove, pMove = pMove->pNext)
{
if (pMove->lConsumerID == lConID)
{
break;
}
}

if (NULL != pMove)
{
/* We found the node to delete. Make the previous (preceding)
node (pPrevNode) point to the node following the node to delete. */
pPrevNode->pNext = pMove->pNext;
if (NULL != pMove->pNext)
{
/* Make the 'pPrev' pointer of the later (succeeding) node point
to the previous (preceding) node of the node to delete. */
(pMove->pNext)->pPrev = pPrevNode;
}
}
}

if (NULL != pMove)
{
/* We found the node to delete. Free the node memory. */
free(pMove);
pMove = NULL;
printf("Node successfully deleted.\n");
}
else
{
printf("No node with the given consumer ID found.\n");
}

} while (0);
}

int main(void)
{
int iOption = 0;

while (3 != iOption)
{
printf("\n1. Add node\n2. Delete node\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

276
 LINKED LISTS

switch (iOption)
{
case 1:
AddNodeSorted();
break;
case 2:
DeleteNode();
break;
default:
break;
}
}

/* Memory cleanup before exit. */

for (PCONSUMPTIONINFO pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
}

return 0;
}

18.2.3 SINGLE CIRCULAR LINKED LIST

The node structure and list construction of single circular linked list is similar to single linear linked list.
The only difference is that, in the case of single circular linked list, the ‘Last/Tail’ node of the list points
back to the ‘First/Head’ node, making the list arrangement circular.

Our program will have two pointers (possibly global), where one will point to the ‘Head’ of the list and
the other to the ‘Tail’. When the list does not have any nodes, both these pointers will point to NULL.
Theoretically speaking, being a circular list, the list does not have any head or tail. But for easy
implementation we consider one node (generally the first added node) as the ‘Head’ of the list and the
node preceding it (generally the last added node) as the ‘Tail’.

When adding a new node to the end of the list, if our list is empty, we mark the new node as both head
and tail of the list. Being a circular linked list, its ‘pNext’ pointer will be made to point to the node itself
(i.e. when we have only a single node in the list). In case of a non-empty list, we make the ‘pNext’ pointer
of the ‘Tail’ node point to the newly created node, and the ‘pNext’ pointer of the new node point to the
‘Head’ node. We mark this new node as the new tail of the list.

Inserting a new node, or deleting an existing node is very similar to a single linear linked list. The only
difference is that while doing these two operations, we should always remember to maintain the
circularity of the list (the Tail node should point to the Head node). In case a node is getting inserted

277
Quick & Indepth C With Data Structures

after the ‘Tail’ node, we should mark it as the new ‘Tail’ node and in case the ‘Tail’ node is getting
deleted, we should make the node situated before the current ‘Tail’ node as the new ‘Tail’ node.

In our next program we will use a single circular linked list. In this program, we will not add the nodes in
sorted order, but will add the new nodes at the tail/end of the list. Being a circular linked list, the list has
no 'Tail'. By 'Tail', we mean the position after the last added node.

#include <stdio.h>
#include <stdlib.h>

typedef struct _CONSUMPTIONINFO

{
long lConsumerID;
int iUnitsConsumed;

struct _CONSUMPTIONINFO *pNext;

}CONSUMPTIONINFO, *PCONSUMPTIONINFO;

PCONSUMPTIONINFO pHead = NULL;

PCONSUMPTIONINFO pTail = NULL;

/* The function creates a new node and adds it to the end of the list i.e. after
the previous 'Tail' node. The newly added node now becomes the new 'Tail' node. */
void AddNode(void)
{
CONSUMPTIONINFO *pNewNode = NULL;

do
{
/* Allocate the new node. */
pNewNode = (CONSUMPTIONINFO *)malloc(sizeof(CONSUMPTIONINFO));
if (NULL == pNewNode)
{
printf("Failed to allocate memory for the new node.\n");
break;
}

/* Accept the consumer details from the user. */

printf("Enter the consumer ID: ");
scanf("%ld", &pNewNode->lConsumerID);

printf("Enter the total units consumed: ");

scanf("%d", &pNewNode->iUnitsConsumed);

if (NULL == pHead)
{
/* Being the only node in the list, our next pointer will point to
ourselves. The new node will be marked as both 'Head' and 'Tail'. */
pNewNode->pNext = pNewNode;
pHead = pTail = pNewNode;
}
else
{
/* Insert the new node between the 'Tail' node and the 'Head' node. */
pNewNode->pNext = pHead;
pTail->pNext = pNewNode;
/* Mark this node as the new 'Tail' node. */
pTail = pNewNode;
}

278
 LINKED LISTS

printf("Node successfully added\n");

} while (0);
}

/* The function accepts the consumer ID and deletes the first

occurring node with that consumer ID. */
void DeleteNode(void)
{
CONSUMPTIONINFO *pPrevNode = NULL;
CONSUMPTIONINFO *pMove = NULL;
long lConID = -1;

do
{
if (NULL == pHead)
{
break;
}

/* Accept the consumer ID to delete. */

printf("Enter the consumer ID to delete: ");
scanf("%ld", &lConID);

/* Start searching from the 'Head' node for the given consumer ID. We will
keep searching until we are back pointing to the 'Head' node again.
'pPrevNode' is the node which precedes the node 'pMove'.
--> |pPrevNode| --> |pMove| ->
Being a circular list, the 'Tail' node points to the 'Head' node. */
pPrevNode = pTail;
pMove = pHead;
do
{
if (pMove->lConsumerID == lConID)
{
/* We found the node to delete. */
if (pHead == pTail)
{
/* Our 'Head' and 'Tail' nodes are the same. This means that we
have only one node in our list and we are about to delete
that one too. So, we will be left with no nodes in the list. */
pHead = pTail = NULL;
}
else if (pMove == pHead)
{
/* We are about to delete our 'Head' node. So, now we need to
adjust our 'Head' pointer to the next node in the list. */
pHead = pHead->pNext;
}
else if (pMove == pTail)
{
/* We are about to delete our 'Tail' node. So, the node before
the 'Tail' node now becomes the new 'Tail' node. */
pTail = pPrevNode;
}

/* We are about to delete the node 'pMove'. So, make the next pointer
of the previous node point to the node which comes after 'pMove'.
If the node to delete (pMove) is |7|, make the next pointer of
the node |4| point to the node |8|.
Previously: --> |3| --> |4| --> |7| --> |8| --> |9| ->
Updated: --> |3| --> |4| --> |8| --> |9| --> */

279
Quick & Indepth C With Data Structures

pPrevNode->pNext = pMove->pNext;

free(pMove);
pMove = NULL;
printf("Node successfully deleted.\n");

break;
}

/* Go to the next node in the list. */

pPrevNode = pMove;
pMove = pMove->pNext;

} while (pMove != pHead);

if (NULL != pMove)
{
printf("No node with the given consumer ID found.\n");
}

} while (0);
}

void TraverseList(void)
{
CONSUMPTIONINFO *pMove = pHead;

if (NULL == pHead)
{
printf("List is empty. No items to show.\n");
return;
}

do
{
printf("Consumer ID: %ld, Units Consumed: %d\n",
pMove->lConsumerID, pMove->iUnitsConsumed);

/* Go to the next node in the list. */

pMove = pMove->pNext;

} while (pMove != pHead);

}

int main(void)
{
int iOption = 0;

while (4 != iOption)
{
printf("\n1. Add node\n2. Delete node\n3. Traverse list\n4. Exit");
printf("\n\nEnter your option (1-4): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
AddNode();
break;
case 2:
DeleteNode();
break;

280
 LINKED LISTS

case 3:
TraverseList();
break;
default:
break;
}
}

/* Memory cleanup before exit. */

for (PCONSUMPTIONINFO pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
}

return 0;
}

18.2.4 DOUBLE CIRCULAR LINKED LIST

The node structure and list construction of a double circular linked list is similar to a double linear
linked list. The only difference is that, in the case of double circular linked list, the ‘Tail’ and ‘Head’
nodes are connected to each other, making the list arrangement circular.

Just like single circular linked list, we will have two pointers (possibly global), one pointing to the ‘Head’
and the other pointing to the ‘Tail’ of the list.

When adding a new node to the end of the list, if our list is empty, we mark the new node as both head
and tail of the list. Being a double circular linked list, both ‘pPrev’ and ‘pNext’ pointers will be made to
point to the node itself (i.e. when we have only a single node in the list). In case of a non-empty list, we
make the ‘pNext’ pointer of the ‘Tail’ node point to the newly created node, ‘pPrev’ pointer of the new
node point to the ‘Tail’ node, the ‘pNext’ pointer of the new node point to the ‘Head’ node, and the
‘pPrev’ pointer of the ‘Head’ node point to the new node. We mark this new node as the new tail of the
list.

Inserting a new node, or deleting an existing node is very similar to a double linear linked list. The only
difference is that while doing these two operations, we should always remember to maintain the
circularity of the list. In case a node is getting inserted after the ‘Tail’ node, we should mark it as the
new ‘Tail’ node and in case the ‘Tail’ node is getting deleted, we should make the node situated before
the current ‘Tail’ node as the new ‘Tail’ node.

In our next program we will use a double circular linked list. The 'TraverseList' function will be exactly

281
Quick & Indepth C With Data Structures

similar to the previous program (for single circular linked list), and so we are not reproducing the same
over here.

#include <stdio.h>
#include <stdlib.h>

typedef struct _CONSUMPTIONINFO

{
long lConsumerID;
int iUnitsConsumed;

struct _CONSUMPTIONINFO *pPrev;

struct _CONSUMPTIONINFO *pNext;

}CONSUMPTIONINFO, *PCONSUMPTIONINFO;

PCONSUMPTIONINFO pHead = NULL;

PCONSUMPTIONINFO pTail = NULL;

/* The function creates a new node and adds it to the end of the list i.e. after the
previous 'Tail' node. The newly added node now becomes the new 'Tail' node. */
void AddNode(void)
{
CONSUMPTIONINFO *pNewNode = NULL;

do
{
/* Allocate the new node. */
pNewNode = (CONSUMPTIONINFO *)malloc(sizeof(CONSUMPTIONINFO));
if (NULL == pNewNode)
{
printf("Failed to allocate memory for the new node.\n");
break;
}

/* Accept the consumer details from the user. */

printf("Enter the consumer ID: ");
scanf("%ld", &pNewNode->lConsumerID);

printf("Enter the total units consumed: ");

scanf("%d", &pNewNode->iUnitsConsumed);

if (NULL == pHead)
{
/* Being the only node in the list, our next and prev pointer will point
to ourselves. The new node will be marked as both 'Head' and 'Tail'. */
pNewNode->pNext = pNewNode;
pNewNode->pPrev = pNewNode;
pHead = pTail = pNewNode;
}
else
{
/* Insert the new node between the 'Tail' node and the 'Head' node. */
pNewNode->pNext = pHead;
pNewNode->pPrev = pTail;
pHead->pPrev = pNewNode;
pTail->pNext = pNewNode;
/* Mark this node as the new 'Tail' node. */
pTail = pNewNode;
}

printf("Node successfully added\n");

282
 LINKED LISTS

} while (0);
}

/* The function accepts the consumer ID and deletes the first

occurring node with that consumer ID. */
void DeleteNode(void)
{
CONSUMPTIONINFO *pMove = NULL;
long lConID = -1;

do
{
if (NULL == pHead)
{
break;
}

/* Accept the consumer ID to delete. */

printf("Enter the consumer ID to delete: ");
scanf("%ld", &lConID);

/* Start searching from the 'Head' node for the given consumer ID. We will
keep searching until we are back pointing to the 'Head' node again. */
pMove = pHead;
do
{
if (pMove->lConsumerID == lConID)
{
/* We found the node to delete. */
if (pHead == pTail)
{
/* Our 'Head' and 'Tail' nodes are the same. This means that we
have only one node in our list and we are about to delete
that one too. So, we will be left with no nodes in the list. */
pHead = pTail = NULL;
}
else if (pMove == pHead)
{
/* We are about to delete our 'Head' node. So, now we need to
adjust our 'Head' pointer to the next node in the list. */
pHead = pHead->pNext;
}
else if (pMove == pTail)
{
/* We are about to delete our 'Tail' node. So, the node before
the 'Tail' node now becomes the new 'Tail' node. */
pTail = pTail->pPrev;
}

/* We are about to delete the node 'pMove'. So, make the next pointer
of the previous node point to the node which comes after 'pMove'.
If the node to delete (pMove) is |7|, make the next pointer of
the node |4| point to the node |8|.
Previously: --> |3| --> |4| --> |7| --> |8| --> |9| -->
Updated: --> |3| --> |4| --> |8| --> |9| --> */
(pMove->pPrev)->pNext = pMove->pNext;
/* Similarly, make the prev pointer of the next node point to the node
which comes before 'pMove'. If the node to delete (pMove) is |7|,
make the prev pointer of the node |8| point to the node |4|.
Previously: <-- |3| <-- |4| <-- |7| <-- |8| <-- |9| <--
Updated: <-- |3| <-- |4| <-- |8| <-- |9| <-- */
(pMove->pNext)->pPrev = pMove->pPrev;

283
Quick & Indepth C With Data Structures

free(pMove);
pMove = NULL;
printf("Node successfully deleted.\n");

break;
}

/* Go to the next node in the list. */

pMove = pMove->pNext;

} while (pMove != pHead);

if (NULL != pMove)
{
printf("No node with the given consumer ID found.\n");
}

} while (0);
}

int main(void)
{
int iOption = 0;

while (3 != iOption)
{
printf("\n1. Add node\n2. Delete node\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
AddNode();
break;
case 2:
DeleteNode();
break;
default:
break;
}
}

/* Memory cleanup before exit. */

for (PCONSUMPTIONINFO pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
}

return 0;
}

284
 STACKS AND QUEUES

STACKS AND QUEUES

[CHAPTER-19]

19.1 INTRODUCTION
Stacks and Queues are two different storage and retrieval mechanisms in data structures. The choice of
usage of a mechanism depends on the problem in hand. Both the mechanisms can be implemented
using arrays and linked lists in C.

19.2 STACKS
A stack is an ordered list in which all insertion and deletion of elements are made at one end only, called
the top. As a real life example, we can visualize this as a stack of books where we need to take the top
items off the stack in order to get things lying under them. A stack is implemented using LIFO (Last In
First Out) arrangement, where the last added element is the first one to be taken out or retrieved.

If the elements 'A', 'B', 'C', 'D', 'E', 'F' are added to the stack (in the given order), then the first element
to be retrieved must be 'F' and the last will be 'A'.

In case of computer programming, the sequence in which nested functions are called and completed
may be visualized as similar to a stack where the last called function completes first. When we are using
recursion, we get a similar stack behavior where the last called instance of the function completes first.
As a matter of fact, the operating system maintains the function call information (function address,
function variables etc.) in a stack arrangement.

The addition of an item to a stack is called push and the retrieval or deletion of an item from the stack is
called pop. Apart from push and pop, we can also perform the following two operations on a stack:
• peek: It returns the item at the top of the stack without removing it from the stack.
• IsEmpty: Checks whether a stack contains any items or not.

We will now write two programs where the first one implements a stack using an integer array and the
second one implements the stack using a single linear linked list.

285
Quick & Indepth C With Data Structures

19.2.1 STACK USING ARRAY

In our next program we will try to implement a stack using an integer array. We will accept numbers
from the user and store them in the array. The entire arrangement of store, display and remove will be
done in LIFO.

#include <stdio.h>

/* Defined values */
#define MAX_STACK_ELEMENTS 10

/* Global variables */
/* The stack array. */
int g_rgiValues[MAX_STACK_ELEMENTS];
/* The variable which keeps track of the number of elements in the stack.
It also helps keep track of the top index of the stack. */
int g_iCurStackCount;

/* Function which adds/pushes a new element on to the stack. */

void PushElement(void)
{
if (MAX_STACK_ELEMENTS == g_iCurStackCount)
{
printf("The stack is full. No more items can be added to the stack.\n");
}
else
{
printf("Enter the value to add: ");
scanf("%d", &g_rgiValues[g_iCurStackCount]);
g_iCurStackCount++;
}
}

/* Function which removes/pops an element out of the stack. */

void PopElement(void)
{
if (0 == g_iCurStackCount)
{
printf("The stack is empty. No items to pop.\n");
}
else
{
g_iCurStackCount--;
printf("Value popped: %d.\n", g_rgiValues[g_iCurStackCount]);
}
}

int main(void)
{
int iOption = 0;

while (3 != iOption)
{
printf("\n1. Add element\n2. Remove element\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{

286
 STACKS AND QUEUES

case 1:
PushElement();
break;
case 2:
PopElement();
break;
default:
break;
}
}

return 0;
}

We are adding the values to the end of the array by pushing the values with increasing array index.
During pop, we are displaying and removing the values starting from the end of the array and
continuing towards the front. This makes the last pushed values get popped out first (LIFO).

19.2.2 STACK USING LINKED LIST

In our next program, we will implement a stack using a single linear linked list. Similar behaviour can
also be achieved using the other types of linked lists.

/* Header Files */
#include <stdio.h>
#include <stdlib.h>

/* Structure Definition */
typedef struct _CONSUMPTIONINFO
{
long lConsumerID;
int iUnitsConsumed;

struct _CONSUMPTIONINFO *pNext;

}CONSUMPTIONINFO, *PCONSUMPTIONINFO;

PCONSUMPTIONINFO pHead = NULL;

/* The function creates a new node and adds it to the start of the list
creating a stack arrangement. It makes the new node the 'Head' of the list
i.e. the last added node becomes the first (Head) node of the list. */
void Push(void)
{
CONSUMPTIONINFO *pNewNode = NULL;

do
{
/* Allocate the new node. */
pNewNode = (CONSUMPTIONINFO *) malloc(sizeof(CONSUMPTIONINFO));
if (NULL == pNewNode)
{
printf("Failed to allocate memory for the new node.\n");
break;
}

/* Accept the consumer details from the user. */

printf("Enter the consumer ID: ");
scanf("%ld", &pNewNode->lConsumerID);

287
Quick & Indepth C With Data Structures

printf("Enter the total units consumed: ");

scanf("%d", &pNewNode->iUnitsConsumed);

pNewNode->pNext = pHead;
pHead = pNewNode;

printf("Node successfully added\n");

} while (0);
}

/* The function pops the first (Head) node from the list i.e. the last
added node is popped first (LIFO). The node following the 'Head' node
becomes the new 'Head' node. */
void Pop(void)
{
CONSUMPTIONINFO *pMove = NULL;

do
{
if (NULL == pHead)
{
printf("The list is empty. Nothing to pop.\n");
break;
}

printf("Consumer ID: %ld, Units Consumed: %d\n",

pHead->lConsumerID, pHead->iUnitsConsumed);

pMove = pHead;
pHead = pHead->pNext;
free(pMove);
pMove = NULL;

printf("Node successfully popped\n");

} while (0);
}

int main(void)
{
int iOption = 0;

while (3 != iOption)
{
printf("\n1. Add node\n2. Delete node\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
Push();
break;
case 2:
Pop();
break;
default:
break;
}
}

288
 STACKS AND QUEUES

/* Memory cleanup before exit. */

for (PCONSUMPTIONINFO pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
}

return 0;
}

In the above program, we are adding the new nodes to the front of the list. The newest node is also
made the ‘Head’ node of the list. We are popping the values from the front of the list producing a stack
arrangement i.e. the last added node is popped first.

19.3 QUEUES
A queue is another ordered list in which insertion and deletion of elements happen at opposite ends. We
can visualize queue as a pipe in which we push items from one end of the pipe and the items come out
from the other end. Another example of a real life approach of queuing systems is whilst waiting in line
at a food store – the first in line will be the first to leave, with new people being added at the end of the
queue. A queue is implemented using FIFO (First In First Out) arrangement, where the first added
element is the first one to be taken out or retrieved. The elements are inserted from one end of the
queue called the rear (also called the tail) and are removed from the other end called the front (also
called the head).

If the elements 'A', 'B', …, ‘J’, ‘K’ are added to the queue (in the given order), then the first element to be
retrieved must be 'A' and the last will be 'K'.

In case of computer programming, the operating system maintains the program information of the
multiple programs running at a time using a queue.

The process of adding an element to a queue is called enqueue and the process of removal of an
element is called dequeue. Apart from enqueue and dequeue, we can also perform the following two
operations on a queue:
• peep: It returns the item at the front of the queue without removing it from the queue.
• IsEmpty: Checks whether a queue contains any items or not.

19.3.1 QUEUE USING ARRAY

Our first example on queue will be to implement the queue using an array. If we add (enqueue) values to
the end of the array and keep removing (dequeue) the values from the front (start) of the array, we will
soon reach the top of the array (the maximum array index), even if we have free space (array elements
whose values have been dequeued/removed) at the front of the array.

289
Quick & Indepth C With Data Structures

To solve this problem, we can resort to enqueue and dequeue of the values using a circular
arrangement where new values are enqueued again from the front (after we reach the array top), if we
have available space at the front of the array.

#include <stdio.h>

/* Defined values */
#define MAX_QUEUE_ELEMENTS 10

/* Global variables */
/* The queue array. */
int g_rgiValues[MAX_QUEUE_ELEMENTS];
/* Variable which keeps track of the start index of the queue i.e. it
keeps track of the first queue element. In simpler words it points to
the index of the element which can next be dequeued. */
int g_iStart = 0;
/* Variable which keeps track of the end index of the queue i.e. it points
to the last enqueued element in the array. */
int g_iEnd = -1;
/* Variable which tracks the total numbers of elements currently in the queue. */
int g_iTotalElements = 0;

/* Function which adds/enqueues a new element to the queue. */

void EnqueueElement(void)
{
if (MAX_QUEUE_ELEMENTS == g_iTotalElements)
{
printf("The queue is full. No more items can be added.\n");
}
else
{
if (MAX_QUEUE_ELEMENTS == (g_iEnd + 1))
{
/* We have reached the array's highest index. Now, wrap around. */
g_iEnd = -1;
}

g_iEnd++;
g_iTotalElements++;
printf("Enter the value to add: ");
scanf("%d", &g_rgiValues[g_iEnd]);
}
}

/* Function which removes/dequeues an element from the queue. */

void DequeueElement(void)
{
if (0 == g_iTotalElements)
{
printf("The queue is empty. No items to dequeue.\n");
}
else
{
printf("Value dequeued: %d.\n", g_rgiValues[g_iStart]);

290
 STACKS AND QUEUES

if (MAX_QUEUE_ELEMENTS == (g_iStart + 1))

{
/* We have reached the array's highest index. Now, wrap around. */
g_iStart = -1;
}

g_iStart++;
g_iTotalElements--;
}
}

int main(void)
{
int iOption = 0;

while (3 != iOption)
{
printf("\n1. Add element\n2. Remove element\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
EnqueueElement();
break;
case 2:
DequeueElement();
break;
default:
break;
}
}

return 0;
}

'g_iStart' moves forward when items are dequeued. As more items are dequeued 'g_iStart' moves
further forward, creating a vacuum at the front of the array. After 'g_iEnd' reaches
'MAX_QUEUE_ELEMENTS', it starts occupying the vacuum left behind at the front of the array when
values were dequeued. This way both 'g_iStart' and 'g_iEnd' move in a circular fashion. Assume, we
have added the values 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 in the given sequence. No more items can be enqueued
as the queue has reached its maximum capacity. We then dequeue the values 1, 2, and 3. This means
our 'g_iStart' has moved to the value 4 (array index 3), creating a free space for 3 new elements at the
front of the array. We can then again resume adding new elements to the queue from the front of the
array. So, 'g_iEnd' loops back to the front of the array. If we add new values 11, 12 and 13, they will
then be placed at the array indices 0, 1 and 2 respectively. 'g_iStart' will then point to 3 (array index)
and 'g_iEnd' will point to 2 (array index).

19.3.2 QUEUE USING LINKED LIST

We will now implement a queue using a single linear linked list in our next program. We will have two
pointers in our program, one pointing to the start of the list ‘Head’ and another pointing to the end
‘Tail’. We will enqueue values to the end of the list and dequeue from the front resulting in a FIFO

291
Quick & Indepth C With Data Structures

behavior.

#include <stdio.h>
#include <stdlib.h>

typedef struct _CONSUMPTIONINFO

{
long lConsumerID;
int iUnitsConsumed;

struct _CONSUMPTIONINFO *pNext;

}CONSUMPTIONINFO, *PCONSUMPTIONINFO;

PCONSUMPTIONINFO pHead = NULL;

PCONSUMPTIONINFO pTail = NULL;

/* The function creates a new node and adds it to the end of the list
creating a queue arrangement. It makes the new node the 'Tail' of the list. */
void Enqueue(void)
{
CONSUMPTIONINFO *pNewNode = NULL;

do
{
/* Allocate the new node. */
pNewNode = (CONSUMPTIONINFO *) malloc(sizeof(CONSUMPTIONINFO));
if (NULL == pNewNode)
{
printf("Failed to allocate memory for the new node.\n");
break;
}

/* Accept the consumer details from the user. */

printf("Enter the consumer ID: ");
scanf("%ld", &pNewNode->lConsumerID);

printf("Enter the total units consumed: ");

scanf("%d", &pNewNode->iUnitsConsumed);

pNewNode->pNext = NULL;

if (NULL == pHead)
{
/* This is the only node in the list. */
pHead = pTail = pNewNode;
}
else
{
/* Make the previous 'Tail' node point to this new node. */
pTail->pNext = pNewNode;
/* Make this node as the new 'Tail' node of the list. */
pTail = pNewNode;
}

printf("Node successfully added.\n");

} while (0);
}

/* The function dequeues the first (Head) node from the list i.e. the first added node is
popped first (FIFO). The node following the 'Head' node becomes the new 'Head' node. */

292
 STACKS AND QUEUES

void Dequeue(void)
{
CONSUMPTIONINFO *pMove = NULL;

do
{
if (NULL == pHead)
{
printf("The list is empty. Nothing to dequeue.\n");
break;
}

printf("Consumer ID: %ld, Units Consumed: %d\n",

pHead->lConsumerID, pHead->iUnitsConsumed);

pMove = pHead;
pHead = pHead->pNext;
free(pMove);
pMove = NULL;

printf("Node successfully dequeued.\n");

} while (0);
}

int main(void)
{
int iOption = 0;

while (3 != iOption)
{
printf("\n1. Add node\n2. Delete node\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
Enqueue();
break;
case 2:
Dequeue();
break;
default:
break;
}
}

/* Memory cleanup before exit. */

for (PCONSUMPTIONINFO pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
}

return 0;
}

293
Quick & Indepth C With Data Structures

TREES

[CHAPTER-20]

20.1 INTRODUCTION
One of the disadvantages of using an unsorted array or linked list to store data is the time needed to
search an item. Since both are linear arrangements, the time required to search an item is proportional
to the size of the array or list. To overcome such problems, we have another kind of data structure
called ‘Tree’. A tree is a non-linear data structure where the stored data is organized hierarchically with
each node (called the parent node) capable of pointing to two or more nodes (called the child nodes),
and the child nodes may optionally point back to the parent node. The structure of a tree is similar to
that of an inverted real life tree which is grounded via the root.

A tree may have zero or more nodes. If it has no nodes, the tree is called a null or empty tree.

The topmost node in the tree is called the Root node. There is only one root node in a tree. The nodes
which do not have any child nodes are called Leaf or External nodes and are at the bottom of the tree.
All the nodes which come in between the root node and the leaf nodes are called the Intermediary or
Internal nodes.

A node is called the parent of another node if it directly points to the other node. Conversely, a node is
called the child of another node (parent) if it is directly pointed to by the other node. In the above
diagram, the node with the value 36 is the parent of the nodes with values 32 and 43 (child nodes).
Nodes which have the same parent are called the siblings.

A node is called a descendant of another node if it can be reached by repeated iteration from the other
node. In the above diagram the node with the value 43 is a descendant of the node 28 but not of the
node 63. A subtree of a tree is a tree consisting of a node 'A' and all the descendants of the node 'A'. So
every node in a tree contains a subtree of itself and all its descendants with the node as the root of that

294
 TREES

subtree. The subtree corresponding to the root node is the entire tree itself. In our example tree, the
node with key value 36 is the root of the subtree with nodes 32, 36 and 43. The root of the entire tree is
the node with key value 47. The number of subtrees of a node is called its degree.

The number of connections in-between a node and the root node of a tree is called the level of a node,
which is equal to ‘1+’ the number of connections. So, the root node itself is at level 1.

The connection between one node and another is called the edge. The number of edges on the longest
path between a node and a leaf is called its height. In the above diagram, the height of the node with
value 47 is 3 and that of 36 is 1. The height of a tree is equal to the height of its root node.

The depth of a node is the number of edges between the node and the root of the tree. The node with
value 63 has a depth of 1.

In this chapter we will discuss about two types of trees – Binary Search Trees and AVL Trees.

20.2 BINARY SEARCH TREES

A binary tree is a tree data structure where each node has at most two children, which are referred to
as the left child and the right child of the node. The tree shown in section 20.1 is an example of a binary
tree. Binary search trees are special type of binary trees where the nodes are arranged according to
their values. They are also called ordered or sorted binary trees as their nodes appear in a sorted order.

A binary search tree (BST in short) allows quick search, addition and removal of items. BST arranges its
nodes in sorted order, so that searches can use the principle of binary search.

Binary search is a search algorithm that finds a target value within a sorted sequence of values. Binary
search compares the target value to the middle element of the sequence. If unequal, half of the
sequence in which the target cannot lie is eliminated and the search continues on the other half. This
process continues until the target value is found or its possibility of existence in the sequence is
eliminated.

In case of BST, the entire binary tree is arranged in such a way which makes the nodes with key values
smaller than the root node lie to the left of the root node, and nodes with key values larger than the root
node lie to the right of it. This principle is followed for every node in the tree, making the entire tree
sorted. To summarize, we may define a BST as a binary tree which has the following properties:
• The left subtree of a node contains only nodes with key values less than the node’s key value.
• The right subtree of a node contains only nodes with key values greater than the node’s key
value.
• The left and right subtrees too must be binary search trees themselves.
• There must be no two nodes with the same key value.

Consider the next BST:

295
Quick & Indepth C With Data Structures

We start searching for a given value from the root node in BST and continue downwards in either the
left subtree or the right subtree depending on the value to search. We compare the search value with
the key value of a node. If same, we return the node. If the search value is smaller than the node's key
value, we recur down the left subtree of the node. Otherwise we recur down its right subtree.

Assume, we are searching for the value 54 in the above BST. We will start from the root node of the tree
and compare its key value with the search value. As the search value is less than the root node's key
value, we will move down the left subtree of the root node. Now, we will be at the node with value '51'.
As the search value is greater than the node's value, we will move down its right subtree, getting us to
the node with value '54' (our desired node).

In a BST, we can traverse the nodes in four different ways:

• Inorder traversal: For every node, all nodes present in its left subtree is traversed first, followed
by that node itself and then by the nodes in its right subtree. This traversal technique produces
the output in sorted order on the node key values.
• Preorder traversal: For every node, its own value is printed first, followed by all nodes present
in its left subtree, and then by the nodes in its right subtree.
• Postorder traversal: For every node, all nodes present in its left branch is traversed first,
followed by the nodes in its right branch, and then that node itself.
• Level order traversal: The nodes are traversed level by level starting from the root node to the
leaf nodes. For each level, the nodes are traversed from left to right. This is an inefficient
traversal technique in terms of space and time requirements, and hence rarely used.

Now, let us write our first binary search tree. The program allows the user to add nodes, delete nodes,
search values, and perform inorder, preorder and postorder traversal.

/* Header Files */
#include <stdio.h>
#include <stdlib.h>

/* Structure Definition */
typedef struct _NODE
{
/* This node's value. */
int iNodeValue;
/* The pointer which points to the left child of this node. */
struct _NODE *pLeftChild;
/* The pointer which points to the right child of this node. */
struct _NODE *pRightChild;

} NODE;

296
 TREES

/* Function which creates and returns a new node. */

NODE *CreateNode(int iValue)
{
/* Allocate the node. */
NODE * pNode = (NODE *) malloc(sizeof(NODE));
if (pNode)
{
pNode->iNodeValue = iValue;
pNode->pLeftChild = pNode->pRightChild = NULL;
}

return pNode;
}

/* Searches the BST for the node with the given value. */
NODE *SearchValue(NODE *pNode, int iValue)
{
do
{
if (NULL == pNode)
{
/* The value was not found. */
printf("The value %d could not be found.\n", iValue);
break;
}

if (pNode->iNodeValue == iValue)
{
/* Found the node with the given value. */
printf("Found the node with value %d.\n", pNode->iNodeValue);
break;
}
else if (iValue < pNode->iNodeValue)
{
/* The value is smaller than this node's value. So, the value
should come in the left branch of this node. */
pNode = SearchValue(pNode->pLeftChild, iValue);
}
else
{
/* The value is larger than this node's value. So, the value
should come in the right branch of this node. */
pNode = SearchValue(pNode->pRightChild, iValue);
}

} while(0);

return pNode;
}

/* Function which performs in-order traversal. For every node, all nodes present in its
left branch is traversed first, followed by that node itself and then by the nodes in
its right branch. This traversal technique always prints the nodes in sorted order.
LEFT CHILD, NODE, RIGHT CHILD. */
void InOrderTraversal(NODE *pNode)
{
if (NULL != pNode)
{
/* Traverse the left branch of this node. All nodes present in the left branch
of this node has been traversed when the (first) below recursive call
completes for this node. This means all values smaller than this node's
value has been printed when this recursive call ends. */
InOrderTraversal(pNode->pLeftChild);

297
Quick & Indepth C With Data Structures

/* The left branch of this node has been printed. Now, print this node's value. */
printf("%d, ", pNode->iNodeValue);

/* Traverse the right branch of this node. All nodes present in the right branch
of this node has been traversed when the below recursive call completes. */
InOrderTraversal(pNode->pRightChild);
}
}

/* Function which performs pre-order traversal. For every node, its own value is
printed first, followed by all nodes present in its left branch, and then by
the nodes in its right branch. NODE, LEFT CHILD, RIGHT CHILD. */
void PreOrderTraversal(NODE *pNode)
{
if (NULL != pNode)
{
/* First, print this node's value. */
printf("%d, ", pNode->iNodeValue);

/* Now, traverse the left branch of this node. */

PreOrderTraversal(pNode->pLeftChild);

/* Now the right branch. */

PreOrderTraversal(pNode->pRightChild);
}
}

/* Function which performs post-order traversal. For every node, all nodes
present in its left branch is traversed first, followed by the nodes in its
right branch, and then that node itself. LEFT CHILD, RIGHT CHILD, NODE. */
void PostOrderTraversal(NODE *pNode)
{
if (NULL != pNode)
{
/* First, traverse the left branch of this node. */
PostOrderTraversal(pNode->pLeftChild);

/* Now the right branch. */

PostOrderTraversal(pNode->pRightChild);

/* Left and right branches of this node has been traversed.

Now, print this node's value. */
printf("%d, ", pNode->iNodeValue);
}
}

/* Function which adds a new node with given value in the BST.
A new value is always added at a leaf position and is never inserted
in-between nodes. The new node always becomes a new leaf node. We start
comparing values from root till we hit a leaf position. Once a leaf position
is found, the new node is added at that position, connecting it to its parent. */
NODE *AddNode(NODE *pNode, int iValue)
{
if (NULL == pNode)
{
/* We have found the position where this node needs to be added. This node
will now become a new leaf node. We will create the new node and its
address will be returned and then assigned to the left/right pointer
(depending on the below 'if' condition check) of the parent node. */

return CreateNode(iValue);
}

298
 TREES

/* Check if the value is smaller than this node's value. If smaller, the value
should come in the left branch of this node. We will start moving down the
left branch until we find the position where the value needs to be added.
Similarly, if the value is larger, we will find the value's position in the
right branch of this node. If this is the node under which the new node
should come, the next recursive call will create the new node and return its
address. We will then assign the returned address to the left/right pointer
(depending on the successful 'if' check below) of this node. */
if (iValue < pNode->iNodeValue)
{
pNode->pLeftChild = AddNode(pNode->pLeftChild, iValue);
}
else if (iValue > pNode->iNodeValue)
{
pNode->pRightChild = AddNode(pNode->pRightChild, iValue);
}
else
{
printf("Duplicate value not allowed.\n");
}

/* This is an existing node. Return its (unchanged) node address. */

return pNode;
}

/* Finds the node with the minimum value under the given 'pNode' subtree. */
NODE *FindMinValueNode(NODE *pNode)
{
/* Start from the given node and continue down its left branch. */
NODE *pCurrent = pNode;

if (pCurrent)
{
/* Loop down to find the leftmost leaf which will have the minimum value. */
while (NULL != pCurrent->pLeftChild)
{
pCurrent = pCurrent->pLeftChild;
}
}

return pCurrent;
}

/* Deletes the node with the given value. If the value is less than the node's
value, we move down the left branch in search of the value. If larger, we move
down the right branch. If the value is not found, the function returns NULL.
If a node's value matches the given value, we de-link the node from the tree.
While de-linking, we may face the following scenarios:
1. The node is a leaf node: Delete the node and return NULL. The parent of the
deleted leaf node starts pointing to NULL in place of the leaf node.
2. The node has only one child node: Delete the node and return the address of
the child node of this node. This makes the child node take the position of
the deleted node.
3. The node has two child nodes: If the node to delete is 'A', find the node with
the lowest value in the right subtree of 'A' (the node to delete). The lowest
value will be the left-most leaf. Assuming 'B' is the leaf node with the lowest
value in the right sub-tree of 'A', replace the value of 'A' with that of 'B'.
Now, delete the node 'B' as done in #1. So, in effect 'B' replaces 'A'. */
NODE *DeleteNode(NODE *pNode, int iValue)
{
NODE *pTemp = NULL;

299
Quick & Indepth C With Data Structures

do
{
if (NULL == pNode)
{
/* The given value does not exist in the tree. */
printf("The value %d does not exist.\n", iValue);
break;
}

/* Check if the given value is smaller than this node's value. If smaller,
the value should be in the left branch/subtree of this node. If larger,
the value should be in the right branch. If same, we delete the node. */
if (iValue < pNode->iNodeValue)
{
/* Look in the left branch/subtree of this node. */
pNode->pLeftChild = DeleteNode(pNode->pLeftChild, iValue);
}
else if (iValue > pNode->iNodeValue)
{
/* Look in the right branch/subtree of this node. */
pNode->pRightChild = DeleteNode(pNode->pRightChild, iValue);
}
else
{
/* We found the node with the given value. */
if (NULL == pNode->pLeftChild)
{
/* The node either has no child nodes or only right child node.
Delete this node and make its child node take its position. */
pTemp = pNode->pRightChild;
free(pNode);
pNode = pTemp;
break;
}
else if (pNode->pRightChild == NULL)
{
/* The node has only one child node and it is the left child.
Delete this node and make its child node take its position. */
pTemp = pNode->pLeftChild;
free(pNode);
pNode = pTemp;
break;
}

/* The node has two child nodes. Find the node with the lowest value in
the right subtree of this node. Assuming this node to be 'A', we try to
find a node 'B' with the lowest value in the right subtree of 'A'. */
pTemp = FindMinValueNode(pNode->pRightChild);

/* Replace this node's value with that of the node with the
lowest value in its right sub-tree. */
pNode->iNodeValue = pTemp->iNodeValue;

/* We have shifted the value/data of 'B' to the node 'A' essentially,

replacing the value of 'A'. Now, delete the node 'B' as it has now
become a duplicate and no longer required. */
pNode->pRightChild = DeleteNode(pNode->pRightChild, pTemp->iNodeValue);
}

} while (0);

return pNode;
}

300
 TREES

/* Deletes the entire tree. */

void DeleteTree(NODE *pNode)
{
if (NULL == pNode)
{
return;
}

/* Delete the left subtree of this node. */

DeleteTree(pNode->pLeftChild);
/* Delete the right subtree of this node. */
DeleteTree(pNode->pRightChild);

/* Now delete this node itself. */

free(pNode);
pNode = NULL;
}

int main(void)
{
int iOption = 0;
int iValue = 0;
NODE *pRoot = NULL;

while (7 != iOption)
{
printf("\n1. Add node");
printf("\n2. Remove node");
printf("\n3. Inorder traversal");
printf("\n4. Preorder traversal");
printf("\n5. Postorder traversal");
printf("\n6. Search value");
printf("\n7. Exit");
printf("\n\nEnter your option (1-7): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
printf("Enter the value to add: ");
scanf("%d", &iValue);
if (NULL == pRoot)
{
pRoot = AddNode(pRoot, iValue);
}
else
{
AddNode(pRoot, iValue);
}
break;

case 2:
printf("Enter the value to delete: ");
scanf("%d", &iValue);
pRoot = DeleteNode(pRoot, iValue);
break;

case 3:
InOrderTraversal(pRoot);
break;

301
Quick & Indepth C With Data Structures

case 4:
PreOrderTraversal(pRoot);
break;

case 5:
PostOrderTraversal(pRoot);
break;

case 6:
printf("Enter the value to search: ");
scanf("%d", &iValue);
SearchValue(pRoot, iValue);
break;

default:
break;
}

printf("\n");
}

/* Clean up by deleting the entire tree. */

DeleteTree(pRoot);
pRoot = NULL;

return 0;
}

20.3 AVL TREES

An AVL tree (named after inventors Adelson, Velski and Landis) is a self-balancing binary search tree. By
self-balancing we mean that the heights of the two child subtrees of any node differ by at most one. If
at any moment (during node addition or deletion) the height difference becomes more than 1, re-
balancing is done to restore this property. Re-balancing is done by rearrangement or rotation of the
nodes of the tree starting from the location of the change (addition or deletion) till the root node. The
difference between the heights of the left subtree and the right subtree is called the Balance Factor.
The balance factor of every node within an AVL tree must be one of the three values -1, 0 or 1.

Lets check out the below binary search trees. The first one is an unbalanced BST and contains a balance
factor outside the values -1, 0 or 1. The second tree is a balanced BST and contains a favorable balance
factor of 0 for each node. Both the trees contain the exact same nodes but arranged a little differently.
The arrangement of the nodes make the first one unbalanced and second one balanced.

Assume, we require to search for the value 71 in both the above trees. In the first tree, we will need to

302
 TREES

traverse through the nodes 75, 51, 54 and 62 to reach our desired node containing the value. In the
second tree, we will require to traverse through the nodes 62 and 75 only to reach our desired node. For
applications that require to do a lot of searching, using a balanced BST may improve the application
efficiency significantly.

Search efficiency may increase many-fold for a balanced BST, because the concept of binary search can
be effectively used. In an unbalanced BST a branch may become so elongated that searching a node
may become almost equivalent to a sequential search. This kind of a situation kills the very purpose of
using a BST. This is where an AVL tree comes to the rescue which guarantees that the tree always
remains balanced.

20.3.1 AVL ROTATIONS

To balance itself, an AVL tree may perform the following four kinds of rotations:
• Right rotation: An AVL tree may become unbalanced, if a new node is inserted in the left
subtree of an existing node. In such a case the tree needs a right rotation. Right rotation is also
desired when an existing node is deleted from the right subtree. The re-balancing of the tree
needs to be done starting from the position of unbalance to the tree root. In the below image,
the node ‘C’ has been added in the left subtree of a node ‘B’. This makes the parent node ‘A’ of
node ‘B’ get unbalanced with a balance factor of 2 (fig-1). This is because ‘A’ was already
having only one child node ‘B’ (no right subtree), when the node ‘C’ was added to the left of
node 'B'. When this situation arises, we need to do a right rotation as depicted in the below
image (fig-2, fig-3).

• Left rotation: An AVL tree may become unbalanced, if a new node is inserted in the right
subtree of an existing node. In such a case the tree needs a left rotation. Left rotation is also
desired when an existing node is deleted from left subtree. The re-balancing of the tree needs to
be done starting from the position of unbalance to the tree root. In the below image, the node
‘C’ has been added in the right subtree of node ‘B’ which makes the parent node ‘A’ of node ‘B’
get unbalanced with a balance factor of -2 (fig-1). This is because ‘A’ was already having only
one child node ‘B’ (no left subtree), when the node ‘C’ was added to the right of node 'B'. When
this situation arises, we do a left rotation as depicted in the next image (fig-2, fig-3).

303
Quick & Indepth C With Data Structures

• Right-Left rotation: We need a double rotation when a situation as in fig-1 of the below image
arises due to a node addition or deletion. Right-Left rotation is a combination of right rotation
followed by a left rotation as shown in the figures 2 to 5.

• Left-Right rotation: When the situation as shown in fig-1 of the below image arises due to a
node addition or deletion, we need a Left-Right rotation. Left-Right rotation is a combination of
left rotation followed by a right rotation as shown in the figures 2 to 5.

304
 TREES

In our next program, we will create an AVL tree and perform node additions and deletions on it. Being a
BST, the search and traversal methods (inorder, preorder and postorder) are exactly the same as our
previous BST example program. Hence, we have skipped these functions over here.

#include <stdio.h>
#include <stdlib.h>

#ifndef max
#define max(a,b) (((a) > (b)) ? (a) : (b))
#endif

typedef struct _NODE

{
/* This node's value. */
int iNodeValue;
/* The pointer which points to the left child of this node. */
struct _NODE *pLeftChild;
/* The pointer which points to the right child of this node. */
struct _NODE *pRightChild;
/* The height of this node in the AVL tree. */
int iHeight;

} NODE;

/* Function which creates and returns a new node. */

NODE *CreateNode(int iValue)
{
NODE * pNode = (NODE *) malloc(sizeof(NODE));
if (pNode)
{
pNode->iNodeValue = iValue;
pNode->iHeight = 0;
pNode->pLeftChild = pNode->pRightChild = NULL;
}

return pNode;
}

/* Function which returns the height of the given node. */

int GetNodeHeight(NODE *pNode)
{
if (NULL == pNode)
{
return -1;
}

return pNode->iHeight;
}

/* Gets the difference in height of the left branch with that

of the right branch of a given node. */
int GetHeightDiff(NODE *pNode)
{
if (NULL == pNode)
{
return 0;
}

return (GetNodeHeight(pNode->pLeftChild) – GetNodeHeight(pNode->pRightChild));

}

305
Quick & Indepth C With Data Structures

/* Function which right rotates a subtree under the given node 'pNodeA'
('pNodeA' is the root node of that subtree).

*/
NODE *RotateRight(NODE *pNodeA)
{
if (NULL == pNodeA || NULL == pNodeA->pLeftChild)
{
return pNodeA;
}

NODE *pNodeB = pNodeA->pLeftChild;

NODE *pNodeC = pNodeB->pRightChild;

/* 'pNodeB' was smaller than 'pNodeA'. So, 'pNodeA' goes to right of 'pNodeB'.
'pNodeC' was smaller than 'pNodeA' (was in its left branch) but, larger than
'pNodeB'. So, it goes to the right branch of 'pNodeB' but left of 'pNodeA'.
Right of 'pNodeB' was previously 'pNodeC', but now its 'pNodeA'.
Left of 'pNodeA' was previously 'pNodeB', but now its 'pNodeC'. */
pNodeB->pRightChild = pNodeA;
pNodeA->pLeftChild = pNodeC;

/* Update the heights of the readjusted nodes. */

pNodeA->iHeight = max(GetNodeHeight(pNodeA->pLeftChild),
GetNodeHeight(pNodeA->pRightChild)) + 1;

pNodeB->iHeight = max(GetNodeHeight(pNodeB->pLeftChild),
GetNodeHeight(pNodeB->pRightChild)) + 1;

/* 'pNodeB' is the new root of this subtree. */

return pNodeB;
}

/* Function which left rotates a subtree under the given node 'pNodeA'
('pNodeA' is the root node of that subtree).

*/
NODE *RotateLeft(NODE *pNodeA)
{
if (NULL == pNodeA || NULL == pNodeA->pRightChild)
{
return pNodeA;
}

NODE *pNodeB = pNodeA->pRightChild;

NODE *pNodeC = pNodeB->pLeftChild;

/* 'pNodeB' was larger than 'pNodeA'. So, 'pNodeA' goes to left of 'pNodeB'.
'pNodeC' was larger than 'pNodeA' (was in its right branch) but, smaller than
'pNodeB'. So, it goes to the left branch of 'pNodeB' but right of 'pNodeA'.
Left of 'pNodeB' was previously 'pNodeC', but now its 'pNodeA'.
Right of 'pNodeA' was previously 'pNodeB', but now its 'pNodeC'. */
pNodeB->pLeftChild = pNodeA;

306
 TREES

pNodeA->pRightChild = pNodeC;

/* Update the heights of the readjusted nodes. */

pNodeA->iHeight = max(GetNodeHeight(pNodeA->pLeftChild),
GetNodeHeight(pNodeA->pRightChild)) + 1;

pNodeB->iHeight = max(GetNodeHeight(pNodeB->pLeftChild),
GetNodeHeight(pNodeB->pRightChild)) + 1;

/* 'pNodeB' is the new root of this subtree. */

return pNodeB;
}

/* Checks if a subtree under the node 'pNode' has become unbalanced. If so,
this function rebalances the subtree and returns the new root node for
that subtree (Previous root node for the subtree was 'pNode'). */
NODE *RebalanceSubTree(NODE *pNode)
{
int iBalance = 0;

do
{
if (NULL == pNode)
{
break;
}

/* Update the height of the given node. */

pNode->iHeight = 1 + max(GetNodeHeight(pNode->pLeftChild),
GetNodeHeight(pNode->pRightChild));

/* Check if the subtree under this node has become unbalanced or not. */
iBalance = GetHeightDiff(pNode);

/* If unbalanced, we will have 4 cases: */

/* Right Rotation Case */

if (iBalance > 1 && 0 <= GetHeightDiff(pNode->pLeftChild))
{
/* The left branch of the subtree is longer. So, we need to rotate
towards right. Rotate and set the new root node of the subtree. */
pNode = RotateRight(pNode);
}
/* Left Rotation Case */
else if (iBalance < -1 && 0 >= GetHeightDiff(pNode->pRightChild))
{
/* The right branch of the subtree is longer. So, we need to rotate
towards left. Rotate and set the new root node of the subtree. */
pNode = RotateLeft(pNode);
}
/* Left-Right Rotation Case */
else if (iBalance > 1 && 0 > GetHeightDiff(pNode->pLeftChild))
{
/* The first left rotation. */
pNode->pLeftChild = RotateLeft(pNode->pLeftChild);
/* Now the right rotation of the main subtree. */
pNode = RotateRight(pNode);
}
/* Right-Left Rotation Case */
else if (iBalance < -1 && 0 < GetHeightDiff(pNode->pRightChild))
{
/* The first right rotation. */
pNode->pRightChild = RotateRight(pNode->pRightChild);

307
Quick & Indepth C With Data Structures

/* Now the left rotation of the main subtree. */

pNode = RotateLeft(pNode);
}

} while (0);

/* Return the new root of the rebalanced subtree. */

return pNode;
}

/* Function which adds a new node with the given value in the AVL tree. The node
addition logic is exactly same as that of BST. Additionally, after the node is
added to the tree, we start rebalancing the tree starting from the parent node
of the newly added node and continuing upwards. */
NODE *AddNode(NODE *pNode, int iValue)
{
if (NULL == pNode)
{
return CreateNode(iValue);
}

if (iValue < pNode->iNodeValue)

{
pNode->pLeftChild = AddNode(pNode->pLeftChild, iValue);
}
else if (iValue > pNode->iNodeValue)
{
pNode->pRightChild = AddNode(pNode->pRightChild, iValue);
}
else
{
printf("Duplicate value not allowed.\n");
return pNode;
}

/* The node has been added. Now rebalance the subtree, if required.
After rebalancing, we return the new root node of this subtree
to the previous recursive call. */
return RebalanceSubTree(pNode);
}

/* Finds the node with the minimum value under the given 'pNode' subtree. */
NODE *FindMinValueNode(NODE *pNode)
{
NODE *pCurrent = pNode;

if (pCurrent)
{
/* Loop down to find the leftmost leaf which will have the minimum value. */
while (NULL != pCurrent->pLeftChild)
{
pCurrent = pCurrent->pLeftChild;
}
}

return pCurrent;
}

/* Deletes the node with the given value. The node deletion logic is very similar
to that of BST. Additionally, after the node is deleted, we start rebalancing
the tree starting from the same level as the deleted node and continuing upwards. */
NODE *DeleteNode(NODE *pNode, int iValue)
{

308
 TREES

NODE *pTemp = NULL;

do
{
if (NULL == pNode)
{
/* The given value does not exist in the tree. */
printf("The value %d does not exist.\n", iValue);
break;
}

if (iValue < pNode->iNodeValue)

{
/* Look in the left branch/subtree of this node. */
pNode->pLeftChild = DeleteNode(pNode->pLeftChild, iValue);
}
else if (iValue > pNode->iNodeValue)
{
/* Look in the right branch/subtree of this node. */
pNode->pRightChild = DeleteNode(pNode->pRightChild, iValue);
}
else
{
/* We found the node with the given value. */
if (NULL == pNode->pLeftChild)
{
/* The node either has no child nodes or only right child node. */
pTemp = pNode->pRightChild;
free(pNode);
pNode = pTemp;
break;
}
else if (pNode->pRightChild == NULL)
{
/* The node has only one child node and it is the left child. */
pTemp = pNode->pLeftChild;
free(pNode);
pNode = pTemp;
break;
}
else
{
/* The node has two child nodes. */
pTemp = FindMinValueNode(pNode->pRightChild);
pNode->iNodeValue = pTemp->iNodeValue;
pNode->pRightChild = DeleteNode(pNode->pRightChild, pTemp->iNodeValue);
}
}

if (NULL == pNode)
{
break;
}

/* If our subtree has become unbalanced, rebalance it. We will continue

rebalancing upwards as we keep coming out of the recursive calls. */
pNode = RebalanceSubTree(pNode);

} while (0);

return pNode;
}

309
Quick & Indepth C With Data Structures

/* Deletes the entire tree. */

void DeleteTree(NODE *pNode)
{
if (NULL == pNode)
{
return;
}

/* Delete the left subtree of this node. */

DeleteTree(pNode->pLeftChild);
/* Delete the right subtree of this node. */
DeleteTree(pNode->pRightChild);

/* Now delete this node itself. */

free(pNode);
pNode = NULL;
}

int main(void)
{
int iOption = 0;
int iValue = 0;
NODE *pRoot = NULL;

while (3 != iOption)
{
printf("\n1. Add node");
printf("\n2. Remove node");
printf("\n3. Exit");
printf("\n\nEnter your option (1-3): ");
scanf("%d", &iOption);
printf("\n");

switch (iOption)
{
case 1:
printf("Enter the value to add: ");
scanf("%d", &iValue);
pRoot = AddNode(pRoot, iValue);
break;

case 2:
printf("Enter the value to delete: ");
scanf("%d", &iValue);
pRoot = DeleteNode(pRoot, iValue);
break;

default:
break;
}

printf("\n");
}

/* Clean up by deleting the entire tree. */

DeleteTree(pRoot);
pRoot = NULL;

return 0;
}

310
 TREES

20.4 SPACE AND TIME COMPLEXITIES

Sometimes, there may be multiple ways to solve a problem. We need to know the performance of the
different algorithms and choose the best one to solve a particular problem. Space and Time
complexities help us analyze the performance of an algorithm.

Space Complexity of an algorithm is the total memory used by the algorithm with respect to its input
size. In simpler words, it is a measure of the amount of memory an algorithm needs. It is stated in terms
of the Big O notation. Big O notation is a mathematical notation that indicates the behavior of a function
when the input tends towards a particular value or infinity. Description of the Big O notation is beyond
the scope of this book.

Time complexity is the computational complexity that estimates the time taken for running an
algorithm. In other words it signifies the total time required by the algorithm to run till its completion. It
is estimated by counting the number of elementary operations performed by the algorithm. Just like
space complexity, time complexity too is stated in terms of the Big O notation.

The space and time complexities of binary search trees are:

Complexities Average Worst
Space O(n) O(n)
Time (Search Node) O(log(n)) O(n)
Time (Add Node) O(log(n)) O(n)
Time (Delete Node) O(log(n)) O(n)

The space and time complexities of AVL trees are:

Complexities Average Worst
Space O(n) O(n)
Time (Search Node) O(log(n)) O(log(n))
Time (Add Node) O(log(n)) O(log(n))
Time (Delete Node) O(log(n)) O(log(n))

20.5 FULL BINARY TREES

A full binary tree is a tree in which every node has either 0 or 2 children. So, every node other than the
leaves have two children. In other words, we will have no nodes containing only one child. A full binary
tree is sometimes called a proper binary tree.

311
Quick & Indepth C With Data Structures

20.6 PERFECT BINARY TREES

A perfect binary tree is a binary tree in which all interior nodes have two children and all leaf nodes are
at the same depth or level. It may be considered as a full binary tree whose leaf nodes are all at the
same depth or level. Hence, all perfect binary trees are full binary trees but, not the other way round.

20.7 COMPLETE BINARY TREES

A complete binary tree is a binary tree in which at every level (except the last), is completely filled, and
all nodes are as far left as possible.

A complete binary tree can also be represented using an array where each element of the array
constitute a node of the tree. If a node is assumed to be at the array element index 'n', its child nodes
will be at the array element indices '2n + 1' (left child) and '2n + 2' (right child) respectively. In the
previous image of the complete binary tree, the numbers beneath each node is the node’s index when
the tree is represented using an array. Hence, the array representation of the tree will look as below:

312
 SORTING AND SEARCHING

SORTING AND SEARCHING

[CHAPTER-21]

21.1 INTRODUCTION
In our previous chapter we learned Binary Search Trees and AVL Trees which help us arrange values in
sorted order. But, these algorithms are limited to be used in tree structures only. If we are dealing with
arrays or linked lists, we need to use algorithms better suited for these data structures. Every sorting
algorithm has affinity towards a specific type of data structure and performs best for that data
structure. For other data structures, the algorithm's performance may degrade substantially. We need
to do careful selection of our sorting algorithm depending on the kind of data structure we are dealing
with. In this chapter we will discuss sorting mechanisms which work on arrays or linked lists.

21.2 QUICKSORT
QuickSort was developed by Tony Hoare in the year 1959. It is a very efficient sorting algorithm and
works with the principle of divide and conquer. It divides/partitions a set of values into two halves and
sorts each half recursively. At each recursive call, it picks a value to act as a pivot (central value) and
arranges all values in such a way so that the values less than the pivot lie to the left of the pivot and
larger values lie to its right. So, all values to the left of the pivot are smaller than all values positioned to
the right of the pivot. We then repeat this process with each of the individual halves recursively and
continue till there are not enough items to partition and sort. This makes each of the individual halves
too getting sorted. The steps can be summarized as below:
• Pick an element, called the pivot, from the set.
• Partition: Rearrange the array so that all elements with values less than the pivot come before
the pivot, while all elements with values greater than the pivot come after it (equal values can
go either way). This makes all values to the left of the pivot to be less than all values positioned
to the right of the pivot.
• Recursively apply the above steps to the two separate sub-sets (one lying to the left of the pivot
and another to the right).

Its best to explain the logic using a simple example. Assume we have an array with five elements 7, 5, 2,
9 and 8. We do the following steps:
1. Take 7 as the pivot.
2. Partition the values. After partitioning we get 5, 2, 7, 9, 8.
3. Sort the first set 5, 2.
a) Take 5 as the pivot.
b) Partition the values. After partitioning we get 2, 5.
4. The array has now become 2, 5, 7, 9, 8.
5. Sort the second set 9, 8.
a) Take 9 as the pivot.

313
Quick & Indepth C With Data Structures

b) Partition the values. After partitioning we get 8, 9.

6. The array has now become 2, 5, 7, 8, 9.

Quicksort is best suited to work with arrays, though we may use it to sort linked lists too. But, its best
performance is achieved when working with arrays.

Let us now implement the algorithm using two programs. The first program will sort an array using the
Quicksort algorithm and the second will do the same on a linked list.

PROGRAM 1: QUICKSORT AN ARRAY

During each recursive call we work within two logical boundaries ‘iLow’ and ‘iHigh’ corresponding to two
array indices. The array section between ‘iLow’ and ‘iHigh’ consists of the elements which need to be
partitioned and sorted. At every recursive call, we first partition the array into two halves - ‘iLow’ to
‘iPivot’ and ‘iPivot + 1’ to ‘iHigh’.

While partitioning, we start with the value of the ‘iLow’ index element as the pivot value. We take two
counters where the first one moves from left to right (increments starting from ‘iLow’) and the second
one moves from right to left (decrements starting from ‘iHigh’) until they cross. All left positioned values
(traversed using the first counter) which are greater than or equal to the pivot value are swapped with
all right positioned values which are smaller than the pivot value (traversed using the second counter).
This makes lesser values to take the lower indices and the higher values take higher indices. We then
return a middle value within the array section as the final pivot value.

Following the partitioning, we get two partitioned halves - ‘iLow’ to ‘iPivot’ and ‘iPivot + 1’ to ‘iHigh’. We
recursively sort these two individual halves again. In every recursive call we repeat the same process of
dividing the array section into two halves and sorting.

#include <stdio.h>

/* The partition needs to be done from 'iLow' to 'iHigh'. We take 'prgiValues[iLow]'

as the starting pivot value. We take two counters, one starting from 'iLow' and
incrementing, and another starting from 'iHigh' and decrementing. We continue until
both these counters cross, the moment when all the values have been partitioned. */
int Partition(int *prgiValues, int iLow, int iHigh)
{
int iPivot = prgiValues[iLow];
int i = iLow – 1;
int j = iHigh + 1;
int iTemp = 0;

while (1)
{
do
{
i++;

} while (i <= j && prgiValues[i] < iPivot);

do
{

314
 SORTING AND SEARCHING

j--;

} while (j >= i && prgiValues[j] > iPivot);

if (i >= j)
{
break;
}

/* Here, we have 'prgiValues[i] >= iPivot' and 'prgiValues[j] <= iPivot'.

But, we should actually have 'prgiValues[i]' less than 'iPivot' and
'prgiValues[j]' greater than 'iPivot'.
Hence, swap the values of these two locations. */
iTemp = prgiValues[i];
prgiValues[i] = prgiValues[j];
prgiValues[j] = iTemp;
}

return j;
}

void Quicksort(int *prgiValues, int iLow, int iHigh)

{
int iPivot = 0;

if (iLow < iHigh)

{
/* Partition the array into two halves. */
iPivot = Partition(prgiValues, iLow, iHigh);
/* Sort the first partitioned half. */
Quicksort(prgiValues, iLow, iPivot);
/* Sort the second partitioned half. */
Quicksort(prgiValues, iPivot + 1, iHigh);
}
}

int main(void)
{
int rgiValues[] = { 23, 28, 78, 62, 11, 5, 68, 92, 34, 51, 15, 85, 47 };
int iArrayNumElements = sizeof(rgiValues) / sizeof(rgiValues[0]);
int idx = 0;

Quicksort(rgiValues, 0, iArrayNumElements – 1);

printf("The array after sorting: \n");

for (; idx < iArrayNumElements; idx++)
{
printf("%d, ", rgiValues[idx]);
}
printf("\n\n");

return 0;
}

PROGRAM 2: QUICKSORT A LINKED LIST

The next program has a little variation from the previous program. As we are working with linked list,
rather than values (as in the case of arrays), the nodes themselves are rearranged due to sorting. This
may lead to a change in the head and tail nodes. The previous head and tail nodes may not be in their
same positions within the list. So, we need to update the original 'Head' and 'Tail' pointers of the caller

315
Quick & Indepth C With Data Structures

function to the new 'Head' and 'Tail' nodes respectively upon completion of the sort.

Within each recursive call, we are working with a section of the list at a time. We are partitioning the list
into two, and individually sorting the two partitions. But during the entire operation, both the partitions
continue to remain as part of the main list. When a list section is partitioned, the nodes get rearranged
and the start and end nodes of that section may change to some other nodes. So, the link between the
starting node (of the section) and the node preceding it (within the main list) will need to be re-
established. Similarly, the link between the last node of a list section and the node succeeding it will
need to be re-established.

#include <stdio.h>
#include <stdlib.h>

/* Defined Values */
#define MAX_ITEMS 10

typedef struct _NODE

{
int iNodeValue; /* This node's value. */
struct _NODE *pNext; /* The pointer which points to the next node. */
struct _NODE *pPrev; /* The pointer which points to the previous node. */

} NODE;

/* In this function we will take the start node as the pivot. All nodes with values
less than the pivot node will be placed before the pivot node, and all nodes with
values greater or equal will be placed after the pivot node. We start traversing
the nodes starting from the node following the pivot node to the end node. We stop
when we have gone past the end node. During traversing, we compare each node's
value with the pivot value. All nodes with values greater than the pivot are left
unchanged (as they are already positioned after the pivot node). All nodes with
values less than the pivot are re-positioned before the pivot node. When a node
with lesser value is found, it is put at the start of this list section and made
the new start node of this section. After the function completes we may have a new
start node and an end node as the nodes may get re-arranged (the position of the
previous end node may change or nodes may come before the previous start node). So
on completion, we update the 'ppStart' and 'ppEnd' pointers with the new addresses. */
NODE *Partition(NODE **ppStart, NODE **ppEnd)
{
NODE *pPivot = *ppStart; /* We take the start node as the pivot */
NODE *pPrevNode = *ppStart;
NODE *pMove = pPrevNode->pNext; /* We start from the node after the start */
NODE *pEndMarker = (*ppEnd)->pNext; /* We partition till we reach this location. */

while (pMove != pEndMarker)

{

316
 SORTING AND SEARCHING

if (pMove->iNodeValue < pPivot->iNodeValue)

{
/* This node's value is less than the pivot. Position it before the pivot. */
/* Disconnect 'pMove' from current location. */
pPrevNode->pNext = pMove->pNext;
if (NULL != pMove->pNext)
{
pMove->pNext->pPrev = pPrevNode;
}

/* Connect 'pMove' to the start of the list section and

mark it as the new 'start' of the list. */
pMove->pPrev = (*ppStart)->pPrev;
(*ppStart)->pPrev = pMove;
pMove->pNext = *ppStart;
*ppStart = pMove;

/* Make 'pMove' again point to the location from

where we will resume traversing. */
pMove = pPrevNode->pNext;
}
else
{
/* 'pPrevNode' always points to the node preceding 'pMove'. */
pPrevNode = pMove;
pMove = pMove->pNext;
}
}

/* Update the new end of the list. */

*ppEnd = pPrevNode;

return pPivot;
}

/* The function accepts the start (ppStart) and end (ppEnd) nodes of the list
which needs to be sorted. It also accepts the node address (in pPreStart)
which precedes the start node. After partitioning, our start and end nodes
may change for which we need to re-link 'pPreStart' with the new start node.
After completion, the pointers 'ppStart' and 'ppEnd' are updated with the
addresses of the new start and end nodes respectively. */
void Quicksort(NODE *pPreStart, NODE **ppStart, NODE **ppEnd)
{
NODE *pPivot = NULL;
NODE *pPivotNext = NULL;

if (NULL != *ppStart && NULL != *ppEnd && *ppStart != *ppEnd)

{
/* Partition the list into two. */
pPivot = Partition(ppStart, ppEnd);
if (NULL != pPreStart)
{
/* The start node might have changed. */
pPreStart->pNext = *ppStart;
}

/* Sort the first partitioned half. */

Quicksort(pPreStart, ppStart, &pPivot);
pPivotNext = pPivot->pNext;
/* Sort the next partitioned half. */
Quicksort(pPivot, &pPivotNext, ppEnd);
}
}

317
Quick & Indepth C With Data Structures

int main(void)
{
NODE *pHead = NULL;
NODE *pTail = NULL;
NODE *pNew = NULL;
NODE *pMove = NULL;

int iCount = 0;

/* Create the linked list. */

for (; iCount < MAX_ITEMS; iCount++)
{
pNew = (NODE *)malloc(sizeof(NODE));
if (NULL == pNew)
{
break;
}

printf("Enter the node's value: ");

scanf("%d", &pNew->iNodeValue);

if (NULL == pHead)
{
pHead = pNew;
pNew->pPrev = NULL;
}
else
{
pTail->pNext = pNew;
pNew->pPrev = pTail;
}

pTail = pNew;
pTail->pNext = NULL;
}

if (NULL != pHead)
{
Quicksort(NULL, &pHead, &pTail);

printf("\nThe list after sorting:\n");

for (pMove = pHead; NULL != pMove; pMove = pMove->pNext)
{
printf("%d, ", pMove->iNodeValue);
}

/* Cleanup the list before exit. */

for (pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
pMove = NULL;
}

printf("\n\n");
}

return 0;
}

318
 SORTING AND SEARCHING

21.3 MERGE SORT

Like Quicksort, Merge Sort is also a divide and conquer algorithm. It was invented by John von
Neumann in the year 1945. It divides a list of items in two halves, recurs itself to sort the two halves
individually and then merges the two sorted halves. It breaks down a list of items into several sublists
until each sublist consists of a single element. It then combines/merges the smaller sorted lists in a way
which produces a sorted merged-list. A merge sort works as follows:
• Divide the unsorted list into n sublists, with each containing just one element.
• Merge these sublists (two sublists at a time) to produce a new sorted sublist. Continue doing so
until there is only one sublist remaining. This will be the sorted list.

The process is better explained in the below image. Assume, we are sorting an array 7, 4, 3, 9, 8, 6, 5.

The entire array is divided into two sub-sections ‘7, 4, 3, 9’ and ‘8, 6, 5’ respectively. The sub-section
‘7, 4, 3, 9’ is further divided into two sub-sections ‘7, 4’ and ‘3, 9’. The sub-section ‘7, 4’ is again
divided into two sub-sections ‘7’ and ‘4’, after which no further divisions are possible. We now merge
the two sub-sections ‘7’, ‘4’ to produce the sorted sequence ‘4, 7’. Similarly, the sub-section ‘3, 9’ is
divided into two sub-sections ‘3’ and ‘9’, and then merged back together to produce ‘3, 9’. The sorted
sections ‘4, 7’ and ‘3, 9’ are now merged together to produce the sorted sequence ‘3, 4, 7, 9’. We now
move towards the right sub-section (‘8, 6, 5’) of the main array and similarly divide and sort it to
produce the sequence ‘5, 6, 8’. We then merge the two sub-sections ‘3, 4, 7, 9’ and ‘5, 6, 8’ to produce
our final sorted result as ‘3, 4, 5, 6, 7, 8, 9’.

Merge sort can not only be used on arrays but also is very effective for linked lists. Unlike Quicksort,
which is naturally inclined towards arrays, merge sort is somewhat inclined more towards linked lists.
For arrays merge sort requires a similar sized array (as the array to be sorted) to work as a temporary
cache while performing the sort. For linked lists, there is no such requirement and hence, its space
complexity is better than merge sort on arrays.

319
Quick & Indepth C With Data Structures

Now let us use two programs to understand the merge sort algorithm. The first program performs
merge sort on an array and the second program performs merge sort on a linked list.

PROGRAM 1: MERGE SORT AN ARRAY

#include <stdio.h>

void Merge(int *prgiSource, int *prgiTarget, int idxBegin, int idxMiddle, int idxEnd)
{
int i = idxBegin;
int j = idxMiddle + 1;
int k = idxBegin;

/* Both the sections (idxBegin to idxMiddle and idxMiddle to idxEnd) are already
in sorted order from previous recursive calls. Now, we scan these two sections
concurrently and copy the next lower value out of these two sections to the
target array. Assume that the two sections have the values as:
Section 1: 17, 23, 26, 42.
Section 2: 25, 29, 37, 45.
We compare the values 17 from section 1 and 25 from section 2. As 17 is lower,
we copy the value to target. Next, we compare the values 23 from section 1
and 25 from section 2. We copy 23 to target array as it is lower. Then we
compare 26 from section 1 with 25 from section 2. We copy 25 to target. We
next compare 26 from section 1 with 29 from section 2. We copy 26. Next, we
compare 42 from section 1 with 29 from section 2. We copy 29. Similarly,
we now compare 42 from section 1 with 37 from section 2. The process continues.
When there are no items left in one section, we copy rest of the items in the
other section to the target array.
We get the final output as:
17, 23, 25, 26, 29, 37, 42, 45. */

while (i <= idxMiddle && j <= idxEnd)

{
if (prgiSource[i] <= prgiSource[j])
{
prgiTarget[k] = prgiSource[i];
i++;
}
else
{
prgiTarget[k] = prgiSource[j];
j++;
}

k++;
}

/* Now copy the remaining items in section 1 to the target (if any left). */
for (; i <= idxMiddle; i++, k++)
{
prgiTarget[k] = prgiSource[i];
}

/* Now copy the remaining items in section 2 to the target (if any left). */
for (; j <= idxEnd; j++, k++)
{
prgiTarget[k] = prgiSource[j];
}
}

320
 SORTING AND SEARCHING

void MergeSort(int *prgiSource, int *prgiTarget, int iBegin, int iEnd)

{
int iMiddle = 0;

if (iEnd - iBegin < 1)

{
/* There is just 1 or less item left to be split and sort. This is
the exit criteria of our recursion and we start coming out. */
return;
}

/* Split the set of items from the middle. */

iMiddle = (iBegin + iEnd) / 2;

/* Recursively split and sort the items in the first section. We are
deliberately reversing the positions of the source and target in both
the below recursive calls to 'MergeSort'. We are making the subsequent
recursive calls and getting the sorted output in 'prgiSource' (the
two calls to 'MergeSort'). We are then using the 'prgiSource' as the
source array for the call to the function 'Merge' for joining the two
sorted sections in 'prgiSource'. 'Merge' function then generates the
final sorted output in 'prgiTarget'. */
MergeSort(prgiTarget, prgiSource, iBegin, iMiddle);
/* Recursively split and sort the items in the second section. */
MergeSort(prgiTarget, prgiSource, iMiddle + 1, iEnd);

/* Merge the two sections. */

Merge(prgiSource, prgiTarget, iBegin, iMiddle, iEnd);
}

int main(void)
{
/* The array to sort. */
int rgiValues[] = { 23, 98, 78, 62, 11 };
/* The array which works as a temporary cache during sorting. */
int rgiTemp[sizeof(rgiValues) / sizeof(rgiValues[0])] = { 0 };
int iArrayNumElements = sizeof(rgiValues) / sizeof(rgiValues[0]);
int idx = 0;

/* In Merge sort we need an equal sized buffer (as the array to sort)
to work as temporary cache during sorting. */
for (idx = 0; idx < iArrayNumElements; idx++)
{
/* Copy the entire contents of the array to our buffer. */
rgiTemp[idx] = rgiValues[idx];
}

/* Now, sort the buffer contents back to our original array. */

MergeSort(rgiTemp, rgiValues, 0, iArrayNumElements – 1);

/* Print the array after sorting. */

printf("The array after sorting: \n");
for (idx = 0; idx < iArrayNumElements; idx++)
{
printf("%d, ", rgiValues[idx]);
}

printf("\n\n");

return 0;
}

321
Quick & Indepth C With Data Structures

PROGRAM 2: MERGE SORT A LINKED LIST

#include <stdio.h>
#include <stdlib.h>

typedef struct _NODE

{
/* This node's value. */
int iNodeValue;
/* The pointer which points to the next node. */
struct _NODE *pNext;

} NODE;

/* Defined Values */
#define MAX_ITEMS 10

/* The function merges the two given lists together and returns the 'Head' of
the merged list. 'pStart1' is the 'Head' of the first list to merge and
'pStart2' is the 'Head' of the second list. */
NODE *Merge(NODE *pStart1, NODE *pStart2)
{
NODE *pStartMerged = NULL;
NODE *pMoveMerged = NULL;
NODE *pNodeToAdd = NULL;

do
{
if (NULL == pStart1)
{
/* The first list is empty. So, our merged list will be the same
as the second list, which already is sorted. */
pStartMerged = pStart2;
break;
}
else if (NULL == pStart2)
{
/* The second list is empty. So, our merged list will be the same
as the first list, which already is sorted. */
pStartMerged = pStart1;
break;
}

/* Traverse both the lists and attach the node with the next lower value to
the merged list. The next starting node in both the lists will have the
smallest value in their respective lists (as both the lists are sorted).
We continue merging until any one of the sub-lists becomes empty. */
while (NULL != pStart1 && NULL != pStart2)
{
if (pStart1->iNodeValue <= pStart2->iNodeValue)
{
pNodeToAdd = pStart1;
pStart1 = pStart1->pNext;
}
else
{
pNodeToAdd = pStart2;
pStart2 = pStart2->pNext;
}

322
 SORTING AND SEARCHING

if (NULL == pMoveMerged)
{
pStartMerged = pMoveMerged = pNodeToAdd;
}
else
{
pMoveMerged->pNext = pNodeToAdd;
pMoveMerged = pMoveMerged->pNext;
}

pMoveMerged->pNext = NULL;
}

/* Now move the remaining items in list-1 to the merged list (if any). */
while (NULL != pStart1)
{
pMoveMerged->pNext = pStart1;
pStart1 = pStart1->pNext;
pMoveMerged = pMoveMerged->pNext;
pMoveMerged->pNext = NULL;
}

/* Now move the remaining items in list-2 to the merged list (if any). */
while (NULL != pStart2)
{
pMoveMerged->pNext = pStart2;
pStart2 = pStart2->pNext;
pMoveMerged = pMoveMerged->pNext;
pMoveMerged->pNext = NULL;
}

} while (0);

return pStartMerged;
}

/* Splits the list starting with 'pHead' into two separate lists. The first
separated list constitute front half of the source list, and the second
separated list constitute the second half. We then return the address of
the starting node of the two split lists using the pointers 'ppStart1' and
'ppStart2'. The function uses slow/fast pointer strategy while traversing
the source list for splitting. The slow pointer moves one node at a time
and the fast pointer moves two pointers at a time. */
void SplitList(NODE *pHead, NODE **ppStart1, NODE **ppStart2)
{
NODE *pFast = NULL;
NODE *pSlow = pHead;

do
{
if (NULL == pSlow || NULL == pSlow->pNext)
{
/* We have less than two nodes in the source list. */
*ppStart1 = pHead;
*ppStart2 = NULL;
break;
}

/* 'pFast' pointer starts from one ahead of the 'pSlow' pointer. */

pFast = pSlow->pNext;

/* Advance the slow pointer by one node and the fast pointer by two nodes. */
while (NULL != pFast)

323
Quick & Indepth C With Data Structures

{
pFast = pFast->pNext;
if (NULL != pFast)
{
pSlow = pSlow->pNext;
pFast = pFast->pNext;
}
}

/* The first separated list starts from 'pHead' till 'pSlow'. */

*ppStart1 = pHead;
/* The second separated list starts from 'pSlow->pNext' till end. */
*ppStart2 = pSlow->pNext;
/* The last node of the first separated list must point to NULL marking
its end. The following statement splits the source list into two. */
pSlow->pNext = NULL;

} while (0);
}

/* The 'Head' node of the original list may change after the sort completes. This
is because the previous 'Head' node might not have the smallest value in the list.
The node with the smallest value in the list becomes the new 'Head' node. So, we
need to update the original 'pHead' pointer of the caller function to the new
'Head' node of the list. Hence, we accept the pointer to the 'pHead' pointer as the
function argument which enables us to update the 'Head' pointer of the caller. */
void MergeSort(NODE **ppHead)
{
NODE *pStart1 = NULL;
NODE *pStart2 = NULL;

if (NULL == *ppHead || NULL == (*ppHead)->pNext)

{
/* We have 0 or 1 node in this list. No sorting required. */
return;
}

SplitList(*ppHead, &pStart1, &pStart2);

/* Recursively split and sort the first separated list. */

MergeSort(&pStart1);
/* Recursively split and sort the second separated list. */
MergeSort(&pStart2);

/* Merge the two sections. */

*ppHead = Merge(pStart1, pStart2);
}

int main(void)
{
NODE *pHead = NULL;
NODE *pTail = NULL;
NODE *pNew = NULL;
NODE *pMove = NULL;

int iCount = 0;

/* Create the linked list. */

for (; iCount < MAX_ITEMS; iCount++)
{
pNew = (NODE *)malloc(sizeof(NODE));
if (NULL == pNew)
{

324
 SORTING AND SEARCHING

break;
}

printf("Enter the node's value: ");

scanf("%d", &pNew->iNodeValue);

if (NULL == pHead)
{
pHead = pNew;
}
else
{
pTail->pNext = pNew;
}

pTail = pNew;
pTail->pNext = NULL;
}

if (NULL != pHead)
{
MergeSort(&pHead);

printf("\nThe list after sorting:\n");

for (pMove = pHead; NULL != pMove; pMove = pMove->pNext)
{
printf("%d, ", pMove->iNodeValue);
}

/* Cleanup the list before exit. */

for (pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
pMove = NULL;
}

printf("\n\n");
}

return 0;
}

21.4 BINARY HEAP

Before moving to our next sorting algorithm ‘heap sort’, let us understand the concept of ‘Binary Heap’
which drives the ‘heap sort’ algorithm. A binary heap is a binary tree with two additional constraints:
1. Shape property: A binary heap is a complete binary tree (refer section 20.7).
2. Heap property: A node's value is either greater than or equal to (>=) or less than or equal to
(<=) the value of its child nodes. This property holds true for every node (except leaf nodes) in
the tree.

Binary heap where the parent node’s value is greater than or equal to (>=) the value of its child nodes is
called Max Heap. Similarly, binary heap where the parent node’s value is less than or equal to (<=) the
value of its child nodes is called Min Heap.

325
Quick & Indepth C With Data Structures

21.5 HEAP SORT

Heap sort was invented by J. W. J. Williams in the year 1964. Heap sort divides the values into two
sections – unsorted (heap section) and sorted. It iteratively shrinks the unsorted section by extracting
the largest element and moving that to the sorted section. Heap sort can be performed in place without
requiring additional memory.

Heap sort algorithm works in the following manner for sorting an array in ascending order:
1. Visualize the entire array as a complete binary tree (refer section 20.7).
2. Build a Max Heap from the array values.
3. This makes the largest item to be stored at the root of the heap. Swap it with the last item of the
heap. Post the swap, the last item of the heap will have the largest value within the heap.
4. Mark the last item of the heap as the starting location of the sorted section.
5. Reduce the size of the heap by 1.
6. Build a Max Heap from the remaining values of the heap.
7. Repeat steps 3 to 6 till the size of the heap is greater than 1.

And for descending order, we do the following:

1. Visualize the entire array as a complete binary tree .
2. Build a Min Heap from the array values.
3. This makes the smallest item to be stored at the root of the heap. Swap it with the last item of
the heap. Post the swap, the last item of the heap will have the smallest value within the heap.
4. Mark the last item of the heap as the starting location of the sorted section.
5. Reduce the size of the heap by 1.
6. Build a Min Heap from the remaining values of the heap.
7. Repeat steps 3 to 6 till the size of the heap is greater than 1.

The function which constructs the in-place heap during heap sort is generally named ‘Heapify’. Now, let
us understand the algorithm using a simple example. Assume, we are sorting an array with element
values ‘7, 4, 6, 2, 1, 3, 5’.
Steps Array Elements
After heapifying the entire array 7, 4, 6, 2, 1, 3, 5
After swap [1] 5, 4, 6, 2, 1, 3, 7

326
 SORTING AND SEARCHING

After re-heapifying the heap section [1] 6, 4, 5, 2, 1, 3, 7

After swap [2] 3, 4, 5, 2, 1, 6, 7
After re-heapifying the heap section [2] 5, 4, 3, 2, 1, 6, 7
After swap [3] 1, 4, 3, 2, 5, 6, 7
After re-heapifying the heap section [3] 4, 2, 3, 1, 5, 6, 7
After swap [4] 1, 2, 3, 4, 5, 6, 7
After re-heapifying the heap section [4] 3, 2, 1, 4, 5, 6, 7
After swap [5] 1, 2, 3, 4, 5, 6, 7
After re-heapifying the heap section [5] 2, 1, 3, 4, 5, 6, 7
After swap [6] 1, 2, 3, 4, 5, 6, 7
After re-heapifying the heap section [6] (Final Sorted Result) 1, 2, 3, 4, 5, 6, 7

NOTE: The numbers within third brackets ‘[]’ in column 1 indicate the iteration count. The numbers
stated in bold in column 2 constitute the sorted section within the array.

Heap sort performs best on arrays. Although it is somewhat slower than a well-implemented Quicksort,
it has the advantage of a better worst-case execution time of O(n log n) against O(n 2) of Quicksort.

Given below is a program which will help us understand the algorithm better.

#include <stdio.h>

/* Swaps the values of two variables. */

void SwapValues(int *pValue1, int *pValue2)
{
int iTemp = 0;

if (NULL != pValue1 && NULL != pValue2)

{
iTemp = *pValue1;
*pValue1 = *pValue2;
*pValue2 = iTemp;
}
}

/* If we visualize the array as a binary tree, this function heapifys'

a subtree rooted at array element location 'iRoot'. 'iNumElements'
refers to the number of elements in the subtree under 'iRoot'. */
void Heapify(int *prgiValues, int iNumElements, int iRoot)
{
/* Start with the assumption that we have the largest value at the root. */
int idxLargest = iRoot;

/* Being a complete binary tree, if 'iRoot' is the index of the root element,
it's left child will be at the array location (2 * iRoot + 1) and the right
child will be at the array location (2 * iRoot + 2). */
int idxLeftChild = 2 * iRoot + 1;
int idxRightChild = 2 * iRoot + 2;

/* Check if we have any left child of this node. We do not have any left child

327
Quick & Indepth C With Data Structures

if its supposed index is greater or equal to the number of elements. */

if (idxLeftChild < iNumElements)
{
if (prgiValues[idxLeftChild] > prgiValues[idxLargest])
{
/* The left child's value is greater than the largest value
(Currently, the root element's value). */
idxLargest = idxLeftChild;
}
}

/* Now, do the same for the right child too (as was done for the left). */
if (idxRightChild < iNumElements)
{
if (prgiValues[idxRightChild] > prgiValues[idxLargest])
{
/* The left child's value is greater than the largest value (the larger
of the root element's value and the left child's value). */
idxLargest = idxRightChild;
}
}

if (idxLargest != iRoot)
{
/* The root element does not have the largest value. One of its
child has the largest value. So, we need to swap the two values
to get the largest value at the root. */
SwapValues(&prgiValues[iRoot], &prgiValues[idxLargest]);

/* Our subtree rooted at 'idxLargest' has changed due to the swap.

We need to re-heapify it. */
Heapify(prgiValues, iNumElements, idxLargest);
}
}

/* The main driver function for the sort procedure. 'prgiValues' is the array
to sort and 'iNumElements' is the total number of elements in the array. */
void HeapSort(int *prgiValues, int iNumElements)
{
int idx = 0;

/* Heapify the entire array. If we visualize the array as a complete binary

tree, where each array element acts as a tree's node, the child nodes
of a node/element 'A' will be at the array indices '(2 * Index of A) + 1'
and '(2 * Index of A) + 2'. On doing heapify, the value of every node/element
will be more than the values of its children. The last non-leaf (Internal)
node of the tree will be at the array index (iNumElements / 2 - 1). We start
heapify-ing from the subtree rooted at this node and continue upwards until
we have heapify-ed the entire array. */
for (idx = iNumElements / 2 - 1; idx >= 0; idx--)
{
Heapify(prgiValues, iNumElements, idx);
}

/* The element at the root will have the highest value within the heap. One by one
extract an element (from the root position) of the heap and swap it with the end
of the heap. We then mark this end as the new start position of the sorted
sequence. We continue doing so until there are no elements left in the heap
and all elements have moved to the sorted portion. If the array currently has
the values '7, 6, 4, 5, 2, 1, 3, 8, 9', where '7, 6, 4, 5, 2, 1, 3' is the
heap portion and '8, 9' is the sorted portion, we swap 7 with 3. This makes
our array become '3, 6, 4, 5, 2, 1, 7, 8, 9' where '7, 8, 9' is the new sorted
portion. After the swap, we will need to re-heapify our heap portion which has

328
 SORTING AND SEARCHING

become '3, 6, 4, 5, 2, 1' due to the swap. */

for (idx = iNumElements - 1; idx >= 1; idx--)
{
SwapValues(&prgiValues[0], &prgiValues[idx]);
Heapify(prgiValues, idx, 0);
}
}

int main(void)
{
int rgiValues[] = { 7, 4, 8, 6, 2, 1, 9, 3, 5 };
int iArrayNumElements = sizeof(rgiValues) / sizeof(rgiValues[0]);
int idx = 0;

HeapSort(rgiValues, iArrayNumElements);

printf("The array after sorting: \n");

for (; idx < iArrayNumElements; idx++)
{
printf("%d, ", rgiValues[idx]);
}
printf("\n\n");

return 0;
}

21.6 INSERTION SORT

Insertion sort works by inserting items to a list. When inserting an item to the list, the list is examined
and the correct position of the new item is found (as per the sorted order), and the item is inserted to
that position in the list. Insertion sort works with the assumption that the items currently in the list are
already in sorted order from previous rounds of sorted insertions. Insertion sort removes one element
from the input data, scans the sorted list to find the location it belongs to, and inserts it there. It
repeats until no input elements remain.

Insertion sort is a very simple sorting algorithm but is much less efficient on large lists than more
advanced algorithms such as Quicksort, Merge sort and Heap sort. Insertion sort is efficient when
working with small lists and is effective when we are building the list in sorted order from the ground up
itself. It performs very well when the input data is mostly sorted, and the algorithm has been optimized
for that purpose. Insertion sort does not require any additional memory space while performing the
sort.

Insertion sort performs poorly for arrays because of the fact that element insertions within arrays is a
time-consuming and processor-intensive task. Insertion sort works good with linked lists as node
insertions can be easily performed on linked lists.

329
Quick & Indepth C With Data Structures

The function ‘AddNodeSorted’ within the program listed under section 18.2.1 of chapter 18 performs
insertion sort while inserting new items to the list.

21.7 BUBBLE SORT

Bubble sort is a very simple sorting algorithm. It repeatedly traverses the list to be sorted, compares
each pair of adjacent items and swaps them if they are in the wrong order. The repeated traversal takes
place until no further swaps are needed, which indicates that the list is sorted.

There are few variations of the sorting algorithm achieving varying levels of optimization. In our
approach, we will consider two sections within our array – unsorted and sorted. The sorted section
contains the values which have already been sorted and have been placed at their desired positions,
hence, requiring no further comparisons and sort. We will only work on the unsorted section, which at
the start of the procedure spans the entire array to be sorted. As we keep sorting, the size of the
unsorted section shrinks and the sorted section expands.

At each pass, we will compare adjacent elements until we reach the end of the unsorted section, which
is: 'Number of array elements – Number of elements already sorted'. The number of iterations (Pass) we
do on the unsorted section is equal to the number of elements in the array minus 1. This is because
when all elements except the last have already been placed in their correct positions, the last element is
the only one left (in the unsorted section) and is already in its desired position within the array.

We further optimize the algorithm to detect an already sorted input array, and thus save ourselves from
useless processing. At each pass (traversal of the unsorted section), we check whether we required to
swap any values. If none of the values were swapped during a pass, it means that all values in the
unsorted section are already in sorted order and require no further processing.

Assume, we are sorting the array ‘5, 2, 4, 1, 3’. The algorithm will work as follows:
Pass 1
5, 2, 4, 1, 3 -> 2, 5, 4, 1, 3 [Swapped as 5 > 2].
2, 5, 4, 1, 3 -> 2, 4, 5, 1, 3 [Swapped as 5 > 4].
2, 4, 5, 1, 3 -> 2, 4, 1, 5, 3 [Swapped as 5 > 1].
2, 4, 1, 5, 3 -> 2, 4, 1, 3, 5 [Swapped as 5 > 3].

Pass 2
2, 4, 1, 3, 5 -> 2, 4, 1, 3, 5
2, 4, 1, 3, 5 -> 2, 1, 4, 3, 5 [Swapped as 4 > 1].
2, 1, 4, 3, 5 -> 2, 1, 3, 4, 5 [Swapped as 4 > 3].

Pass 3
2, 1, 3, 4, 5 -> 1, 2, 3, 4, 5 [Swapped as 2 > 1].
1, 2, 3, 4, 5 -> 1, 2, 3, 4, 5

330
 SORTING AND SEARCHING

Pass 4
1, 2, 3, 4, 5 -> 1, 2, 3, 4, 5

Bubble sort works best with arrays and should be avoided with linked lists. Although the algorithm is
simple, it is slow for most input data. It gives a much better performance when the input data set is
small and is mostly in sorted order. The advantage bubble sort has over most other sorting algorithms
(except insertion sort) is the ability to detect that the list is sorted and thus, skip much of the
unnecessary processing.

Let us now write a program which implements the bubble sort algorithm.

#include <stdio.h>

void BubbleSort(int *prgiValues, int iNumElements)

{
if (NULL == prgiValues)
{
return;
}

int i = 0;
int j = 0;
int iTemp = 0;
int iSwapped = 0;

/* 'i' accounts for the passes we do on the unsorted section. We do not need a
pass for the last unsorted item as it is the only one left in the unsorted
section and thus occupying the only available position within the array. */
for (; i < iNumElements - 1; i++)
{
iSwapped = 0;

for (j = 0; j < iNumElements - i - 1; j++)

{
/* At each pass on the unsorted section, we go till the end of the
unsorted section within the array. 'i' is the number of passes
we have already performed and thus equals the number of elements
occupying the sorted section (already sorted). */

/* Compare the adjacent elements and swap if required. */

if (prgiValues[j] > prgiValues[j + 1])
{
iTemp = prgiValues[j];
prgiValues[j] = prgiValues[j + 1];
prgiValues[j + 1] = iTemp;
iSwapped = 1;
}
}

if (0 == iSwapped)
{
/* We did not require to do any swaps. This means all the values in
the unsorted section are already in sorted order and require no
further processing. */
break;
}
}
}

331
Quick & Indepth C With Data Structures

int main(void)
{
int rgiValues[] = { 7, 4, 6, 2, 1, 3, 5 };
int iArrayNumElements = sizeof(rgiValues) / sizeof(rgiValues[0]);
int idx = 0;

BubbleSort(rgiValues, iArrayNumElements);

printf("The array after sorting: \n");

for (; idx < iArrayNumElements; idx++)
{
printf("%d, ", rgiValues[idx]);
}
printf("\n\n");

return 0;
}

21.8 SELECTION SORT

Just like heap sort, selection sort too divides an array into two sections – sorted and unsorted. Initially,
the sorted section is empty and the unsorted section spans the entire array. As the sort progresses,
more values are transferred from the unsorted section to the sorted section making the sorted section
expand and the unsorted section shrink. This continues until there are no values left in the unsorted
section. The sorted section occupies the start of the array (left) and the unsorted section occupies the
end (right).

At each iteration, the algorithm scans the unsorted section of the array to find the element with the
least (or highest in case of descending sort) value. When its found, its value is swapped with the value of
the left-most element of the unsorted section. Post the swap, the sorted section is expanded by one
element (to include the left-most element from the unsorted section) and the unsorted section is
shrinked by one element (to exclude the left-most element).

The steps for sorting an array with 'n' elements in ascending order can be summarized as below:
1. Scan the unsorted section for the element with the least value.
2. Swap its value with the left-most element in the unsorted section.
3. Increment the starting index of the unsorted section.
4. Repeat steps 1 to 3 for 'n-1' times.

Note: For sorting in descending order, we will search for the element with the largest value (instead of
the least value) in step 1.

Assume we are sorting an array 7, 2, 9, 3, 4. The sort operation will be performed in the following steps:
• 7, 2, 9, 3, 4 -> 2, 7, 9, 3, 4 [Unsorted section start index: 0]
• 2, 7, 9, 3, 4 -> 2, 3, 9, 7, 4 [Unsorted section start index: 1]
• 2, 3, 9, 7, 4 -> 2, 3, 4, 7, 9 [Unsorted section start index: 2]
• 2, 3, 4, 7, 9 -> 2, 3, 4, 7, 9 [Unsorted section start index: 3, Final sorted result]

332
 SORTING AND SEARCHING

Though selection sort is a very simple algorithm, it performs poorly in comparison to some other
algorithms like Quicksort, Merge sort and Heap sort. Insertion sort too, in most cases perform better
than selection sort. The Heap sort algorithm, though very similar to the selection sort algorithm,
performs much better due to the use of binary heaps.

Selection sort works best with arrays, though, we may use it on linked lists too. The program-1 in
section 5.5 of chapter 5, performs selection sort on an array. Here, we will write a program which will
perform selection sort on a linked list.

#include <stdio.h>

/* Defined Values */
#define MAX_ITEMS 10

typedef struct _NODE

{
/* This node's value. */
int iNodeValue;
/* The pointer which points to the next node. */
struct _NODE *pNext;

} NODE;

void SelectionSort(NODE *pHead)

{
NODE *pMove1 = NULL;
NODE *pMove2 = NULL;
NODE *pMin = NULL;
int iTemp = 0;

/* 'pMove1' marks the start of the unsorted section.

Initially, it is the entire list. */
for (pMove1 = pHead; NULL != pMove1->pNext; pMove1 = pMove1->pNext)
{
/* Start with the assumption that the start node of the
unsorted section has the least value. If we find a node
with a lesser value, we will mark it as the new 'min'. */
pMin = pMove1;

/* Scan the entire unsorted section and try to find the

node with the least value. */
for (pMove2 = pMove1->pNext; NULL != pMove2; pMove2 = pMove2->pNext)
{
if (pMove2->iNodeValue < pMin->iNodeValue)
{
/* A new 'min' node is found. */
pMin = pMove2;
}
}

if (pMove1 != pMin)
{
/* The 'min' node is not the same as 'pMove1'. Swap their values. */
iTemp = pMin->iNodeValue;
pMin->iNodeValue = pMove1->iNodeValue;
pMove1->iNodeValue = iTemp;
}
}
}

333
Quick & Indepth C With Data Structures

int main(void)
{
NODE *pHead = NULL;
NODE *pTail = NULL;
NODE *pNew = NULL;
NODE *pMove = NULL;

int iCount = 0;

/* Create the linked list. */

for (; iCount < MAX_ITEMS; iCount++)
{
/* Allocate the new node. */
pNew = (NODE *) malloc(sizeof(NODE));
if (NULL == pNew)
{
break;
}

printf("Enter the node's value: ");

scanf("%d", &pNew->iNodeValue);

if (NULL == pHead)
{
/* The list is empty. This is the first node of the list. */
pHead = pNew;
}
else
{
pTail->pNext = pNew;
}

pTail = pNew;
pTail->pNext = NULL;
}

if (NULL != pHead)
{
SelectionSort(pHead);

printf("\nThe list after sorting:\n");

for (pMove = pHead; NULL != pMove; pMove = pMove->pNext)
{
printf("%d, ", pMove->iNodeValue);
}

/* Cleanup the list before exit. */

for (pMove = pHead; NULL != pMove; pMove = pHead)
{
pHead = pHead->pNext;
free(pMove);
pMove = NULL;
}

printf("\n\n");
}

return 0;
}

334
 SORTING AND SEARCHING

21.9 BINARY SEARCH TREE

Just like insertion sort, we could build a BST from the ground up yielding us a sorted list at every stage
of our list. When we are creating the list by adding items to it, we build it as a binary search tree or an
AVL tree. The inorder traversal of the tree will generate the output in sorted order. This strategy helps
us attain a very favourable space and time complexities.

21.10 SPACE AND TIME COMPLEXITIES

The space and time complexities of the various sorting algorithms are:
Algorithm Space Time (Best) Time (Average) Time (Worst)
Merge sort O(n) with arrays and O(n log(n)) O(n log(n)) O(n log(n))
O(log(n)) with linked lists
Quick sort O(n) O(n log(n)) O(n log(n)) O(n2)
Heap sort O(1) O(n log(n)) O(n log(n)) O(n log(n))
Insertion sort O(1) O(n) O(n2) O(n2)
Bubble sort O(1) O(n) O(n2) O(n2)
Selection sort O(1) O(n2) O(n2) O(n2)

21.11 SEARCHING
When searching for an item within a list, we can opt for any of the following methods:
• Linear search
• Binary search
• Binary search tree
• Jump search

21.11.1 LINEAR SEARCH

Linear search is the simplest form of searching which involves searching every element within the list
until the desired element is found. This searching method is also called sequential search. We start the
search from the initial element of the list and continue searching until we have found the desired
element or we have reached the end of the list. Though linear search is an extremely slow search
technique, the advantage it has over most other search techniques is the fact that the input list need
not be in sorted order. Linear search can be performed on almost all types of lists (arrays, linked lists
etc.). The time complexity of a linear search is O(n).

21.11.2 BINARY SEARCH

Binary search works only on a sorted list. It searches a sorted array by repeatedly dividing the search
range in half and testing the middle element. The middle element is returned, if its value equals the
search value. In an ascending ordered list, if the search value is less than the value of the element in the
middle, the search continues in the lower half of the range. If the search value is more, the search
continues in the upper half. We continue doing so until we find the search value or there are no
elements in a range.

335
Quick & Indepth C With Data Structures

Assume, we are searching for a value ‘iSrch’ within a sorted array ‘arr’. The array is sorted in ascending
order and contains ‘N’ number of elements. Two variables ‘iStart’ and ‘iEnd’ are declared, and initialized
with the values 0 (zero) and ‘N-1’ respectively. We look at the flowchart for binary searching such an
array:

The time complexity of binary search is O(log(n)). Not only binary search has a simple implementation
but also, it is very efficient. Let us understand it using a small program.

#include <stdio.h>

int BinarySearch(int *prgiValues,

int iNumElements,
int iSrchValue)
{
int iStart = 0;
int iEnd = iNumElements – 1;
int iMid = 0;
int iItemIndex = -1;

while (iStart <= iEnd)

{
/* Find the middle index between the start and end indices. */
iMid = (iStart + iEnd) / 2;
if (iSrchValue == prgiValues[iMid])
{
/* Found the value. */

336
 SORTING AND SEARCHING

iItemIndex = iMid;
break;
}
else if (iSrchValue < prgiValues[iMid])
{
/* The search value should lie in the lower half. */
iEnd = iMid – 1;
}
else
{
/* The search value should lie in the upper half. */
iStart = iMid + 1;
}
}

return iItemIndex;
}

int main(void)
{
/* 'rgiValues' must be sorted in ascending order. */
int rgiValues[] = { 10, 14, 19, 23, 27, 36, 39, 48, 53, 57, 59, 61 };
int iArrayNumElements = sizeof(rgiValues) / sizeof(rgiValues[0]);
int idx = 0;
int iSrchValue = 23;

/* Now, Search the value. */

idx = BinarySearch(rgiValues, iArrayNumElements, iSrchValue);

if (idx < 0)
{
printf("The value %d could not be found.", iSrchValue);
}
else
{
printf("The value %d is at array index %d.", iSrchValue, idx);
}

printf("\n\n");
return 0;
}

21.11.3 BINARY SEARCH TREE

We may arrange the data as a binary search tree, which will enable us to easily search for an item within
the tree using the very principles of a binary search tree (refer section 20.2). Search is very efficient for
a BST as it uses binary search for the purpose. We can achieve best search performance only if the BST
is balanced like an AVL tree. For a balanced BST, we will achieve a time complexity of O(log(n)).

21.11.4 JUMP SEARCH

Jump search also called block search checks fewer elements (than a linear search) by skipping ahead
some elements. Like binary search, jump search too needs the array to be in sorted order. The algorithm
uses a jump interval to jump a specific number of elements while keeping a track of the position it held
before the jump. The elements spanning the last position to the jump location constitute the current
block. The jump location is also called the jump step or simply step. The algorithm keeps jumping
elements till the value of the jump step is less than the search value (greater in case the array is sorted

337
Quick & Indepth C With Data Structures

in descending order). If the search value is found to be less or equal, it will most likely be present in the
current block. We then perform a linear search in the current block for the value.

Assume we are searching the value ‘17’ within the array 2, 5, 9, 11, 15, 17, 21, 27, 31. We set our jump
interval as 3. We hold the last position (before the jump) in the variable ‘iStart’ and the position after the
jump (jump step) in ‘iStep’. The elements between ‘iStart’ and ‘iStep’ constitute the current block.
• iStart is initialized to 0 (zero) and iStep is initialized to 3 (jump interval).
• iStep (value 11) is less than search value 17. This makes us to continue the jump search.
• iStart becomes 4 and iStep becomes 6.
• iStep (value 21) is more than search value 17. The search value should lie in the current block
(array index 4 to array index 6).
• Perform linear search in the current block (array index 4 to array index 6).

Jump search is an improvement over linear search but, is less efficient than binary search. It is prudent
to use binary search instead of jump search wherever possible. The only advantage of jump search over
binary search is that we generally require lesser number of backward jumps than in binary search. For
jump search, we mostly move forward, and resort to backward traversal only when we are doing linear
search within the identified block. Jump search has a time complexity of O(sqrt(n)).

#include <stdio.h>

int JumpSearch(int *prgiValues,

int iNumElements,
int iSrchValue)
{
int iStart = 0;
int iIncrement = sqrt(iNumElements);
int iStep = iIncrement;
int iItemIndex = -1;

/* Find the block where our search value may be present. Jump search
by skipping 'iIncrement' elements each time till the search value is
more than the step element's value. */
for (; iStep < iNumElements; iStep += iIncrement)
{
if (iSrchValue <= prgiValues[iStep])
{
/* The search value (if present) should lie in this block. */
break;
}

iStart = iStep + 1;
}

if (iStep >= iNumElements)

{
/* We have crossed the end of the array. Re-adjust to the array end. */
iStep = iNumElements – 1;
}

/* We found the block. Perform linear search within this block (moving backwards). */
for (; iStep >= iStart; iStep--)
{
if (iSrchValue > prgiValues[iStep])

338
 SORTING AND SEARCHING

{
/* We crossed the location where the value could have existed. This means
the search value does not exist in the array. Assume we are searching the
value 8 within the block of values 6, 7, 9. We start from the value 9
and move backwards. We will stop at the value 7 when our search value
of 8 becomes greater than the element's value (meaning the search value
does not exist). We perform backward search as it achieves better
performance when a step value is equal to the search value. */
break;
}
else if (iSrchValue == prgiValues[iStep])
{
iItemIndex = iStep;
break;
}
}

return iItemIndex;
}

int main(void)
{
/* 'rgiValues' must be sorted in ascending order. */
int rgiValues[] = { 10, 14, 19, 23, 27, 36, 39, 48, 53, 57, 59, 61 };
int iArrayNumElements = sizeof(rgiValues) / sizeof(rgiValues[0]);
int idx = 0;
int iSrchValue = 57;

/* Now, Search the value. */

idx = JumpSearch(rgiValues, iArrayNumElements, iSrchValue);

if (idx < 0)
{
printf("The value %d could not be found.", iSrchValue);
}
else
{
printf("The value %d is at array index %d.", iSrchValue, idx);
}

printf("\n\n");
return 0;
}

339
Quick & Indepth C With Data Structures

C LIBRARY FUNCTIONS

[APPENDIX-1]
C language has a collection of very powerful library functions declared in multiple header files. From
time to time, new functions are added to existing header files depending on the compiler and the C
version we are working with. Also, new header files are created as per the changing requirements. Here
we will discuss some of the major library functions.

Some of the C header files are:

Header File Description
Contains the assert macro which assists in detecting specific errors in
assert.h
debug versions of a program.
complex.h (since C99) Complex number arithmetic functions.
ctype.h Single character test and conversion functions.
errno.h Macros for testing error codes reported by library functions.
float.h Defines the limits of various float types (in macros).
Defines the limits of various fundamental integral types depending on
limits.h
the architecture and compiler used (in macros).
Defines localization specific settings, such as culture-specific date
locale.h
formats or country-specific currency symbols.
math.h Mathematical functions.
Provides types and macros which enable functions to accept variable
stdarg.h
number of arguments.
stdbool.h (since C99) Contains macros for the Boolean data type.
stddef.h Common macro definitions.
stdint.h (since C99) Fixed-width integer types.
stdio.h Standard input and output functions
Some utility functions for string conversion, memory allocation, random
stdlib.h
number generation etc.
Defines the noreturn macro to identify functions which will not return the
stdnoreturn.h (since C11)
control back to the caller function post their execution.
string.h String handling functions.
time.h Time information and manipulation functions.
wchar.h (since C95) String manipulation functions (wide character versions).
Single character test and conversion functions (wide character
wctype.h (since C95)
versions).

340
 C LIBRARY FUNCTIONS

Note: Many of the functions discussed in this appendix are deprecated and are replaced with their newer
versions. Please refer to chapter-16 for a list of the deprecated functions and their replacements.

assert.h
assert If the argument expression of this macro equal to zero, a message is written to
the standard error device and the program execution is terminated. The message
will include the expression whose assertion failed, the name of the source file
and the line number where it happened.

complex.h
cabs, cabsf, cabsl Computes the complex absolute value.
carg, cargf, cargl Computes argument of a complex number.
cimag, cimagf, cimagl Computes the imaginary part of a complex number.
creal, crealf, creall Computes the real part of a complex number.
conj, conjf, conjl Computes the complex conjugate.
cproj, cprojf, cprojl Computes the complex projection into Riemann sphere.
cexp, cexpf, cexpl Computes the complex exponential.
clog, clogf, clogl Computes the complex logarithm.
csqrt, csqrtf, csqrtl Computes the complex square root.
cpow, cpowf, cpowl Computes the complex power.
csin, csinf, csinl Computes complex sine.
ccos, ccosf, ccosl Computes complex cosine.
ctan, ctanf, ctanl Computes complex tangent.
casin, casinf, casinl Computes complex arc sine.
cacos, cacosf, cacosl Computes complex arc cosine.
catan, catanf, catanl Computes complex arc tangent.
csinh, csinhf, csinhl Computes complex hyperbolic sine.
ccosh, ccoshf, ccoshl Computes complex hyperbolic cosine.
ctanh, ctanhf, ctanhl Computes complex hyperbolic tangent.
casinh, casinhf, casinhl Computes complex hyperbolic arc sine.
cacosh, cacoshf, cacoshl Computes complex hyperbolic arc cosine.
catanh, catanhf, catanhl Computes complex hyperbolic arc tangent.

ctype.h (ANSI)/wctype.h (Wide Character)

ctype.h wctype.h
isalnum iswalnum Checks if the specified character is alphanumeric.
isalpha iswalpha Checks if the specified character is alphabetic
isblank iswblank Checks if the specified character is a blank character.
iscntrl iswcntrl Checks if the specified character is a control character.
isdigit iswdigit Checks if the specified character is a decimal digit.
isgraph iswgraph Checks if the specified character has graphical representation.

341
Quick & Indepth C With Data Structures

ctype.h wctype.h
islower iswlower Checks if the specified character is a lowercase letter.
isprint iswprint Checks if the specified character can be printed.
ispunct iswpunct Checks if the specified character is a punctuation character.
isspace iswspace Checks if the specified character is a white-space character.
isupper iswupper Checks if the specified character is an uppercase letter.
isxdigit iswxdigit Checks if the specified character is a hexadecimal digit.
tolower towlower Returns the lowercase equivalent of the given character.
toupper towupper Returns the uppercase equivalent of the given character.

errno.h
errno An integer variable whose value indicates the last error faced by a standard library
function. Its value can be changed within the program code.

float.h
FLT_RADIX Base used for the floating-point types.
FLT_MANT_DIG Precision of the mantissa. ‘FLT’ is for float data type, ‘DBL’ is for
DBL_MANT_DIG double and ‘LDBL’ is for long double.
LDBL_MANT_DIG
FLT_DIG Number of decimal digits that can be rounded to a floating point and
DBL_DIG back without any change to the value. ‘FLT’ is for float data type, ‘DBL’
LDBL_DIG is for double and ‘LDBL’ is for long double.
FLT_MIN_EXP Minimum value for the exponent that generates a normalized floating
DBL_MIN_EXP point number. ‘FLT’ is for float data type, ‘DBL’ is for double and ‘LDBL’
LDBL_MIN_EXP is for long double.
FLT_MIN_10_EXP Minimum value for the exponent of a Base-10 expression that
DBL_MIN_10_EXP generates a normalized floating point number. ‘FLT’ is for float data
LDBL_MIN_10_EXP type, ‘DBL’ is for double and ‘LDBL’ is for long double.
FLT_MAX_EXP Maximum value for the exponent that generates a normalized floating
DBL_MAX_EXP point number. ‘FLT’ is for float data type, ‘DBL’ is for double and ‘LDBL’
LDBL_MAX_EXP is for long double.
FLT_MAX_10_EXP Maximum value for the exponent of a Base-10 expression that
DBL_MAX_10_EXP generates a normalized floating point number. ‘FLT’ is for float data
LDBL_MAX_10_EXP type, ‘DBL’ is for double and ‘LDBL’ is for long double.
FLT_MAX Maximum value which can be finitely represented as a floating point
DBL_MAX number. ‘FLT’ is for float data type, ‘DBL’ is for double and ‘LDBL’ is for
LDBL_MAX long double.
FLT_MIN Minimum value which can be finitely represented as a floating point
DBL_MIN number. ‘FLT’ is for float data type, ‘DBL’ is for double and ‘LDBL’ is for

342
 C LIBRARY FUNCTIONS

LDBL_MIN long double.

FLT_EPSILON Difference between 1 and the next smallest floating point number.
DBL_EPSILON ‘FLT’ is for float data type, ‘DBL’ is for double and ‘LDBL’ is for long
LDBL_EPSILON double.
FLT_ROUNDS Rounding mode of floating points operations.
-1: Indeterminable
0: Towards zero
1: To the nearest number
2: Towards positive infinity
3: Towards negative infinity

limits.h
CHAR_BIT Number of bits in a char.
SCHAR_MIN Minimum possible value of a signed char.
SCHAR_MAX Maximum possible value of a signed char.
UCHAR_MAX Maximum possible value of an unsigned char.
CHAR_MIN Minimum possible value of a char.
CHAR_MAX Maximum possible value of a char.
MB_LEN_MAX Maximum number of bytes in a multi-byte character, for any locale.
SHRT_MIN Minimum possible value of a short int.
SHRT_MAX Maximum possible value of a short int.
USHRT_MAX Maximum possible value of an unsigned short int.
INT_MIN Minimum possible value of an int.
INT_MAX Maximum possible value of an int.
UINT_MAX Maximum possible value of an unsigned int.
LONG_MIN Minimum possible value of a long int.
LONG_MAX Maximum possible value of a long int.
ULONG_MAX Maximum possible value of an unsigned long int.
LLONG_MIN Minimum possible value of a long long int.
LLONG_MAX Maximum possible value of a long long int.
ULLONG_MAX Maximum possible value of an unsigned long long int.

locale.h
struct lconv Contains formatting info for numeric and monetary values.
setlocale Sets or retrieves the locale information.
localeconv Gets locale formatting information.

343
Quick & Indepth C With Data Structures

math.h
acos Calculates inverse cosine.
asin Calculates inverse sine.
atan Calculates inverse tangent.
atan2 Calculates inverse tangent (two parameters).
ceil Returns the smallest whole number greater than the parameter.
cos Calculates cosine.
cosh Calculates hyperbolic cosine.
exp Calculates the exponential value.
fabs Calculates the absolute value of the floating point number.
floor Returns the largest whole number not greater than the parameter.
fmod Returns the floating point remainder for the division of the parameter1
by parameter2.
frexp Returns mantissa and exponent of the floating point number.
ldexp Multiplies a floating point number by an integral power of two.
log Calculates natural logarithm.
log10 Calculates Base-10 logarithm.
modf Breaks into fractional and integral parts.
pow Returns parameter1 to the power of parameter2.
sin Calculates sine.
sinh Calculates hyperbolic sine.
sqrt Calculates the square root.
tan Calculates tangent.
tanh Calculates hyperbolic tangent.
Below functions are available in C99 and above
acosh Calculates inverse hyperbolic cosine.
asinh Calculates inverse hyperbolic sine.
atanh Calculates inverse hyperbolic tangent.
cbrt, cbrtf, cbrtl Calculates the cube root.
copysign, copysignf, copysignl Returns the value of parameter1 with the sign of parameter2.
erf, erff, erfl Calculates error function.
erfc, erfcf, erfcl Calculates complementary error function.
exp2, exp2f, exp2l Calculates 2 to the power of the parameter.
expm1, expm1f, expm1l Calculates exponential minus 1.
fdim, fdimf, fdiml Returns positive difference.
fma, fmaf, fmal Multiply and add.

344
 C LIBRARY FUNCTIONS

fmax, fmaxf, fmaxl Returns the largest of the parameter1 and parameter2.
fmin, fminf, fminl Returns the smallest of the parameter1 and parameter2.
hypot, hypotf, hypotl Calculates the hypotenuse.
ilogb, ilogbf, ilogbl Calculates unbiased Base-2 exponent.
lgamma, lgammaf, lgammal Calculates natural logarithm of absolute value of the gamma function.
llrint, llrintf, llrintl Rounds to integer using current rounding mode (returns long long).
lrint, lrintf, lrintl Rounds to integer using current rounding mode (returns long).
llround, llroundf, llroundl Rounds to the nearest whole number (returns long long).
lround, lroundf, lroundl Rounds to the nearest whole number (returns long).
log1p, log1pf, log1pl Returns the natural logarithm of the parameter plus 1.
log2, log2f, log2l Calculates the Base-2 logarithm.
logb, logbf, logbl Extracts exponent from the floating point number.
nan Returns a NaN from the provided string argument.
nearbyint, nearbyintf, nearbyintl Rounds the floating point number to nearest whole number.
nextafter, nextafterf, nextafterl Returns next re-presentable value after parameter1 and in the
direction of parameter2.
nexttoward, nexttowardf, Returns next re-presentable value after parameter1 and in the
nexttowardl direction of parameter2 (higher precision).
remainder, remainderf, Computes the remainder of the quotient, rounded to the nearest
remainderl integral value.
remquo, remquof, remquol Computes the remainder of two values, and stores an integer value
with the sign and approximate magnitude of the quotient in a location
specified as a parameter.
rint, rintf, rintl Rounds to integer using the current rounding mode.
round, roundf, roundl Rounds to integer, rounding halfway cases away from zero.
scalbln, scalblnf, scalblnl Scales parameter1 by FLT_RADIX raised to the power of the
parameter2 (where parameter2 is long int).
scalbn, scalbnf, scalbnl Scales parameter1 by FLT_RADIX raised to the power of the
parameter2 (where parameter2 is int).
tgamma, tgammaf, tgammal Returns the gamma function.
trunc, truncf, truncl Rounds towards zero, returning the nearest integral value that is not
larger in magnitude.
Macros
fpclassify Returns the floating-point classification of the parameter.
isfinite Determines whether the argument has a finite value.
isinf Determines whether the argument is infinite.
isnan Checks if the argument is not a number.

345
Quick & Indepth C With Data Structures

signbit Returns whether the argument is negative.

isgreater Checks if parameter1 is greater than parameter2.
isgreaterequal Checks if parameter1 is greater than or equal to parameter2.
isless Checks if parameter1 is less than parameter2.
islessequal Checks if parameter1 is less than or equal to parameter2.
islessgreater Checks if parameter1 is less or greater than parameter2.

stdarg.h
va_list Type for iterating the variable arguments list. Holds the information
required for the macros ‘va_start’, ‘va_arg’ and ‘va_end’ to work.
va_start Initializes a variable argument list.
va_arg Retrieves the next argument in the list.
va_end Marks the end of the use of the variable arguments list.
va_copy (since C99) Copies a variable arguments list to another.

stdbool.h
bool Expands to _Bool, the boolean data type.
true Expands to the value 1.
false Expands to the value 0 (zero).

stddef.h
ptrdiff_t Represents the result of any valid pointer subtraction operation.
size_t Unsigned integral type.
NULL Null pointer.
offsetof Returns the offset value in bytes of the specified member field within
the data structure or union.

stdint.h
Types
int8_t Signed 8-bit integer.
uint8_t Unsigned 8-bit integer.
int16_t Signed 16-bit integer.
uint16_t Unsigned 16-bit integer.
int32_t Signed 32-bit integer.
uint32_t Unsigned 32-bit integer.
int64_t Signed 64-bit integer.
uint64_t Unsigned 64-bit integer.

346
 C LIBRARY FUNCTIONS

intmax_t Signed integer type with maximum capacity.

uintmax_t Unsigned integer type with maximum capacity.
Type ranges
INT8_MIN Minimum possible value of int8_t.
INT8_MAX Maximum possible value of int8_t.
UINT8_MAX Maximum possible value of uint8_t.
INT16_MIN Minimum possible value of int16_t.
INT16_MAX Maximum possible value of int16_t.
UINT16_MAX Maximum possible value of uint16_t.
INT32_MIN Minimum possible value of int32_t.
INT32_MAX Maximum possible value of int32_t.
UINT32_MAX Maximum possible value of uint32_t.
INT64_MIN Minimum possible value of int64_t.
INT64_MAX Maximum possible value of int64_t.
UINT64_MAX Maximum possible value of uint64_t.
INTMAX_MIN Minimum possible value of intmax_t.
INTMAX_MAX Maximum possible value of intmax_t.
UINTMAX_MAX Maximum possible value of uintmax_t.
SIZE_MAX Maximum possible value of size_t.
PTRDIFF_MIN Minimum possible value of ptrdiff_t.
PTRDIFF_MAX Maximum possible value of ptrdiff_t.
SIG_ATOMIC_MIN Minimum possible value of sig_atomic_t.
SIG_ATOMIC_MAX Maximum possible value of sig_atomic_t.
WCHAR_MIN Minimum possible value of wchar_t.
WCHAR_MAX Maximum possible value of wchar_t.
WINT_MIN Minimum possible value of wint_t.
WINT_MAX Maximum possible value of wint_t.

stdio.h/wchar.h (Wide character)

stdio.h wchar.h
File open or close functions
fopen _wfopen Creates or opens a file.
freopen _wfreopen Reopens a file stream with a different file or mode.
fflush Flushes the data in the stream to the output device/disk.
setbuf Sets the buffer for a stream.
setvbuf Changes the stream buffering.

347
Quick & Indepth C With Data Structures

stdio.h wchar.h
fclose Closes an already opened file.
Stream positioning functions
fgetpos Gets current position in the stream.
fsetpos Sets the position indicator of the stream to the new position.
ftell Gets the current position in the stream.
rewind Sets position of the stream to the beginning.
fseek Repositions stream position indicator to the new position.
Single character input or output functions
fgetc fgetwc Reads a character from the stream.
fputc fputwc Writes a character to the stream.
getc getwc Reads a character from the stream.
putc putwc Writes a character to the stream.
ungetc ungetwc Ungets a character from stream. The file position indicator is
repositioned to the location as was before the last ‘get’
operation.
getchar getwchar Gets a character from the standard input device (stdin).
putchar putwchar Writes a character to the standard output device (stdout).
String input or output functions
fgets fgetws Reads a string from the stream.
fputs fputws Writes a string to the stream.
gets _getws Gets a string from the standard input device (stdin).
puts _putws Writes a string to the standard output device (stdout).
Binary mode input or output functions
fread Reads a block of data from the stream.
fwrite Writes a block of data to the stream.
Formatted input or output functions
fprintf fwprintf Writes formatted data to the stream.
fscanf fwscanf Reads formatted data from the stream.
printf wprintf Writes formatted data to the standard output device
(stdout).
scanf wscanf Reads formatted data from the standard input device (stdin).
snprintf (since C99) _snwprintf Writes formatted output to the provided buffer (buffer size
specified by caller).
sprintf swprintf Writes formatted data to the string.
sscanf swscanf Reads formatted data from the specified string.

348
 C LIBRARY FUNCTIONS

stdio.h wchar.h
vfprintf vfwprintf Writes formatted data from variable argument list to the
stream.
vfscanf (since C99) vfwscanf Reads formatted data from the stream into variable
argument list.
vprintf vwprintf Writes formatted data from variable argument list to the
standard output device (stdout).
vscanf (since C99) vwscanf Reads formatted data from standard input device (stdin) into
variable argument list.
vsnprintf (since _vsnwprintf Writes formatted data from variable argument list to the
C99) provided buffer (buffer size specified by caller).
vsprintf vswprintf Writes formatted data from variable argument list to string.
vsscanf (since C99) vswscanf Reads formatted data from string into variable argument list.
File system operation functions
remove _wremove Removes the specified file.
rename _wrename Renames the specified file.
tmpfile Creates a temporary binary file for read/write with a filename
guaranteed to be unique.
tmpnam _wtmpnam Returns a string containing a filename different from any
existing file.
Error handling functions
clearerr Clears all error indicators for the stream.
feof Checks if the end of file indicator for the stream is set.
ferror Checks if the error indicator for the stream is set.
perror Interprets the value in ‘errno’ and prints it to the standard
output error stream.
Macros
BUFSIZ Default buffer size.
EOF End of file marker.
FILENAME_MAX Maximum length of filenames.
FOPEN_MAX Maximum simultaneous open streams.
NULL Null pointer marker.
TMP_MAX Maximum number of temporary files.

Note: The functions with name starting with an underscore ‘_’ are compiler specific and may not be
available on all C compilers.

349
Quick & Indepth C With Data Structures

stdlib.h (ANSI)/wchar.h (Wide character)

stdlib.h wchar.h
Type conversion functions
atof _wtof Converts the given string to double.
atoi _wtoi Converts the given string to integer.
atol _wtol Converts the given string to long integer.
atoll (since C99) _wtoll Converts the given string to long long integer.
strtod wcstod Converts the given string to double. The function also sets
parameter2 to point to the character following the number
within the string.
strtof (since C99) wcstof Converts the given string to float. The function also sets
parameter2 to point to the character following the number
within the string.
strtol wcstol Converts the given string to long integer using the specified
base. The function also sets parameter2 to point to the
character following the number within the string.
strtold (since C99) wcstold Converts the given string to long double. The function also
sets parameter2 to point to the character following the
number within the string.
strtoll (since C99) wcstoll Converts the given string to long long integer using the
specified base. The function also sets parameter2 to point to
the character following the number within the string.
strtoul wcstoul Converts the given string to unsigned long integer using the
specified base. The function also sets parameter2 to point to
the character following the number within the string.
strtoull (since C99) wcstoull Converts the given string to unsigned long long integer using
the specified base. The function also sets parameter2 to point
to the character following the number within the string.
Dynamic memory management functions
malloc Allocates block of memory from heap.
calloc Allocates a zero-initialized memory block from heap.
realloc Re-allocates an existing block of memory.
free Deallocates a memory block.
Mathematical
abs Returns the absolute value.
div Returns the integral quotient and remainder after dividing
parameter1 by parameter2 (integer).
labs Returns the absolute value as long integer.

350
 C LIBRARY FUNCTIONS

stdlib.h wchar.h
ldiv Returns the integral quotient and remainder after dividing
parameter1 by parameter2 (long integer).
llabs (since C99) Returns the absolute value as long long integer.
lldiv (since C99) Returns the integral quotient and remainder after dividing
parameter1 by parameter2 (long long integer).
Pseudo random number generation function
rand Generates random number.
srand Initializes the random number generator using a seed value.
Environment management functions
abort Aborts current process, producing an abnormal program
termination.
atexit Sets the function to be executed when the program
terminates.
exit Terminates current process normally, performing all regular
cleanup before terminating.
getenv _wgetenv Gets the environment string corresponding to the
environment name specified as a parameter to the function.
system _wsystem Executes a system (OS) command.
Searching and sorting functions
bsearch Performs binary search in an array.
qsort Sorts the specified array.
Multibyte/Wide character operation functions
mblen Gets the length of the multibyte character.
mbtowc Converts multibyte sequence to wide character.
wctomb Converts wide character to multibyte sequence.
mbstowcs Converts multibyte string to wide character string.
wcstombs Converts wide character string to multibyte string.
Macros and types
EXIT_FAILURE Program failure termination code.
EXIT_SUCCESS Program success termination code.
MB_CUR_MAX Maximum size of multibyte characters.
NULL Null pointer.
RAND_MAX Maximum value returned by the rand function.
size_t Unsigned integral type.
Note: The functions with name starting with an underscore ‘_’ are compiler specific and may not be
available on all C compilers.

351
Quick & Indepth C With Data Structures

stdnoreturn.h
noreturn (Since C11) Expands to _Noreturn which specifies that the function does not return
to its point of invocation.

string.h (ANSI)/wchar.h (Wide Character)

[For detailed explanations, please refer to section 9.7.1]
string.h wchar.h
Copy functions
memcpy wmemcpy Copies a block of memory from source to destination.
memmove wmemmove Moves a block of memory from source to destination.
strcat wcscat Appends a copy of the source string to the destination string.
strcpy wcscpy Copies the source string to the destination.
strncat wcsncat Appends a copy of the source string to the destination string,
subject to a maximum of 'n' characters.
strncpy wcsncpy Copies the source string to the destination, subject to a maximum
of 'n' characters.
Compare functions
memcmp wmemcmp Compares one block of memory to another.
strcmp wcscmp Compares one string to another.
strcoll wcscoll Compares one string to another using the locale.
strncmp wcsncmp Compares one string to another, subject to a maximum of 'n'
characters.
strxfrm wcsxfrm Copies a string from a source to destination after transforming it
using the current locale.
Search functions
memchr wmemchr Searches a maximum of 'n' characters for the first occurance of the
given character within the given block of memory.
strchr wcschr Searches a string for the first occurance of the given character.
strcspn wcscspn Scans parameter1 (string) for the first occurrence of any of the
characters in parameter2 (string). It returns the index of the first
occurrence.
strpbrk wcspbrk Scans parameter1 (string) for the first occurrence of any of the
characters in parameter2 (string). It returns a pointer to the first
occurrence.
strrchr wcsrchr Searches a string for the last occurance of the given character.
strspn wcsspn Returns the index of the first character in parameter1 (string), that
does not belong to any of the characters in parameter2 (string).

352
 C LIBRARY FUNCTIONS

string.h wchar.h
strstr wcsstr Returns a pointer to the first occurance of parameter2 (string) in
parameter1 (string).
strtok wcstok A sequence of calls to this function splits parameter1 (string) into
tokens, where every token is a sequence of contiguous characters
separated from the next token by any of the delimiters (characters)
specified in parameter2 (string).
Miscelleneous functions
memset wmemset Sets the first 'n' bytes of the block of memory pointed by
parameter1 to the value specified in parameter2.
strerror _wcserror Gets a system error message string corresponding to the given
error code.
strlen wcslen Returns the length of a string.
Note: The functions with name starting with an underscore ‘_’ are compiler specific and may not be
available on all C compilers.

time.h/wchar.h (Wide character)

time.h wchar.h
asctime _wasctime Converts the provided 'tm' structure to a string
clock Returns the processor time (clock ticks) elapsed since the
program started
ctime _wctime Converts the provided ‘time_t’ value to a string
difftime Returns the difference (in seconds) between two times
gmtime Converts the provided ‘time_t’ to ‘tm’ structure as UTC time
localtime Converts the provided ‘time_t’ to ‘tm’ structure as local time
mktime Returns the 'time_t' corresponding to the provided tm structure
strftime wcsftime Formats the time specified as ‘tm’ structure to a string
time Returns the current system time
Types
clock_t Clock tick count
struct tm Structure containing a date and time broken down into its
components
time_t The number of seconds elapsed since Jan 1, 1970, 00:00 hours
UTC
Macros
CLOCKS_PER_SEC The number of clock ticks (clock_t) per second
Note: The functions with name starting with an underscore ‘_’ are compiler specific and may not be
available on all C compilers.

353
Quick & Indepth C With Data Structures

wchar.h
[Apart from the functions and macros already stated alongside other header files (in this appendix),
‘wchar.h’ has the following additional functions and macros]
btowc Returns the wide character equivalent of the specified single byte character.
wctob Converts the wide character to single byte character.
mbrlen Returns the length of the specified multibyte character.
mbsinit Checks if the provided ‘mbstate_t’ is at initial conversion state.
mbrtowc Converts the provided multibyte sequence to a wide character.
mbsrtowcs Converts the provided multibyte string to a wide character string.
wcrtomb Converts the provided wide character to multibyte sequence.
wcsrtombs Converts the provided wide character string to multibyte string.
Types
mbstate_t Type that holds the state information when converting between multibyte
characters and wide characters (both ways).
wchar_t Wide character type.
wint_t Wide integer type to hold error codes too.
Macros
WCHAR_MAX Maximum possible value of wchar_t.
WCHAR_MIN Minimum possible value of wchar_t.
WEOF Wide end of file marker.

Note: Many of the functions discussed in this appendix are deprecated and are replaced with their
newer versions. Please refer to chapter-16 for a list of the deprecated functions and their
replacements.

354