Chapter 9
Uploaded by
Er Anupam DattaChapter 9
Uploaded by
Er Anupam DattaChapter 9
Two Dimensional Arrays
In Chapter 7 we have seen that C provides a compound data structure for storing a list of related
data. For some applications, however, such a structure may not be sucient to capture the
organization of the data. For example, in our payroll task, we have several pieces of information
(hours worked, rate of pay, and regular and over time pay) for a list of employees. We have
organized such data as individual arrays, one for each \column", to form the payroll data base
but conceptually, this information is all related. In this chapter we will introduce a data structure
which allows us to group together all such information into a single structure | a two dimensional
array. For a data base application, we can think of this 2D organization as an array of arrays. As
another example of where such a structure is convenient, consider an array of names. We have
seen that we can store a name in a string, which is an array of characters. Then an array of strings
is also an array of arrays, or a two dimensional array.
In this chapter we will discuss how we can declare two dimensional arrays, and use them in
applications. We will see how we can access the data in such a structure using indices and pointers,
and see how this concept can be extended to multi-dimensional arrays. We will present examples
of 2 dimensional arrays for data base applications, string sorting and searching, and solutions
to systems of simultaneous linear algebraic equations, useful in scientic, engineering, and other
applications, e.g. electronic circuit analysis, economic analysis, structural analysis, etc. The one
restriction in the use of this data type is that all of the data stored in the structure must be of
the same type. (We will see how we can remove this restriction in the next chapter).
9.1 Two Dimensional Arrays
Our rst task is to consider a number of exams for a class of students. The score for each exam is
to be weighted dierently to compute the nal score and grade. For example, the rst exam may
contribute 30% of the nal score, the second may contribute 30%, and the third contribute 40%.
We must compute a weighted average of the scores for each student. The sum of the weights for
all the exams must add up to 1, i.e. 100%. Here is our task:
WTDAVG: Read the exam scores from a le for several exams for a class of students. Read
the percent weight for each of the exams. Compute the weighted average score for each student.
Also, compute the averages of the scores for each exam and for the weighted average scores.
We can think of the exam scores and the weighted average score for a single student as a data
record and and represent it as a row of information. The data records for a number of students,
341
342 CHAPTER 9. TWO DIMENSIONAL ARRAYS
then, is a table of such rows. Here is our conceptual view of this collection of data:
exam1 exam2 exam3 weighted avg
student1 50 70 75 ??
student2 90 95 100 ??
student3 89 87 92 ??
...
studentn 90 90 91 ??
Let us assume that all scores will be stored as integers
even the weighted averages, which will
be computed as oat, will be rounded o and stored as integers. To store this information in a
data structure, we can store each student's data record, a row containing three exam scores and
the weighted average score, in a one dimensional array of integers. The entire table, then, is an
array of these one dimensional arrays | i.e. a two dimensional array. With this data structure,
we can access a record for an individual student by accessing the corresponding row. We can also
access the score for one of the exams or for the weighted average for all students by accessing each
column. The only restriction to using this data structure is that all items in an array must be
of the same data type. If the student id is an integer, we can even include a column for the id
numbers.
Suppose we need to represent id numbers, scores in 3 exams, and weighted average of scores
for 10 students
we need an array of ten data records, one for each student. Each data record
must be an array of ve elements, one for each exam score, one for the weighted average score,
and one for the student id number. Then, we need an array, scores10] that has ten elements
each element of this array is, itself, an array of 5 integer elements. Here is the declaration of an
array of integer arrays:
int scores10]5]
The rst range says the array has ten elements: scores0], scores1], : : : scores9]. The
second range says that each of these ten arrays is an array of ve elements. For example, scores0]
has ve elements: scores0]0], scores0]1], : : : scores0]4]. Similarly, any other element
may be referenced by specifying two appropriate indices, scoresi]j]. The rst array index
references the ith one dimensional array, scoresi]
the second array index references the j th
element in the one dimensional array, scoresi]j].
A two dimensional array lends itself to a visual display in rows and columns. The rst in-
dex represents a row, and the second index represents a column. A visual display of the array,
scores10]5], is shown in Figure 9.1. There are ten rows, (0-9), and ve columns (0-4). An
element is accessed by row and column index. For example, scores2]3] references an integer
element at row index 2 and column index 3.
We will see in the next section that, as with one dimensional arrays, elements of a two dimen-
sional array may be accessed indirectly using pointers. There, we will see the connection between
two dimensional arrays and pointers. For now, we will use array indexing as described above and
remember that arrays are always accessed indirectly. Also, just as with one dimensional arrays, a
2D array name can be used in function calls, and the called function accesses the array indirectly.
We can now easily set down the algorithm for our task:
9.1. TWO DIMENSIONAL ARRAYS 343
col. 0 col. 1 col. 2 col. 3 col. 4
row 0 scores 0] 0] scores 0] 1] scores 0] 2] scores 0] 3] scores 0] 4]
row 1 scores 1] 0] scores 1] 1] scores 1] 2] scores 1] 3] scores 1] 4]
row 2 scores 2] 0] scores 2] 1] scores 2] 2] scores 2] 3] scores 2] 4]
row 3 scores 3] 0] scores 3] 1] scores 3] 2] scores 3] 3] scores 3] 4]
row 4 scores 4] 0] scores 4] 1] scores 4] 2] scores 4] 3] scores 4] 4]
row 5 scores 5] 0] scores 5] 1] scores 5] 2] scores 5] 3] scores 5] 4]
row 6 scores 6] 0] scores 6] 1] scores 6] 2] scores 6] 3] scores 6] 4]
row 7 scores 7] 0] scores 7] 1] scores 7] 2] scores 7] 3] scores 7] 4]
row 8 scores 8] 0] scores 8] 1] scores 8] 2] scores 8] 3] scores 8] 4]
row 9 scores 9] 0] scores 9] 1] scores 9] 2] scores 9] 3] scores 9] 4]
Figure 9.1: Rows and Columns in A Two Dimensional Array
read the number of exams into no_of_exams
get weights for each of the exams
read exam scores and id number for each student
into a two dimensional array
for each student, compute weighted average of scores in the exams
compute average score for each of the exams and
for the weighted average
print results
We can easily write the top level program driver using functions to do the work of reading scores,
getting the weights, computing the weighted averages, printing scores, averaging each set of scores,
and printing the averages. The driver is shown in Figure 9.2.
We have declared an array, scores]], with MAX rows and COLS columns, where these macro
values are large enough to accommodate the expected data. We have used several functions,
which we will soon write and include in the same program le. Their prototypes as well as those
of other functions are declared at the head of the le. In the driver, getwts() reads the weights
for the exams into an array, wts], returning the number of exams. The function, read scores(),
reads the data records into the two dimensional array, scores]], and returns the number of
data records. The function, wtd avg(), computes the weighted averages of all exam scores, and
avg scores() computes an average of each exam score column as well as that of the weighted
average column. Finally, print scores() and print avgs() print the results including the input
data, the weighted averages, and the averages of the exams.
Let us rst write getwts(). It merely reads the weight for each of the exams as shown in Figure
9.3. The function prompts the user for the number of exam scores, and reads the corresponding
number of oat values into the wts] array. Notice that the loop index, i begins with the value
1. This is because the element wts0], corresponding to the student id column, does not have a
weight and should be ignored. After the weights have been read, we ush the keyboard buer of
any remaining white space so that any kind of data (including character data) can be read from
344 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: wtdavg.c
Other Source Files: avg.c
Header Files: avg.h
This program computes weighted averages for a set of exam scores for
several individuals. The program reads scores from a file, computes
weighted averages for each individual, prints out a table of scores,
and prints averages for each of the exams and for the weighted average.
*/
#include <stdio.h>
#define MAX 20
#define COLS 5
int getwts(float wts])
FILE *openinfile(void)
int read_scores(int ex]COLS], int lim, int nexs)
void wtd_avg(int ex]COLS], int lim, int nexs, float wts])
void avg_scores(int ex]COLS], int avg], int lim, int nexs)
void print_scores(int ex]COLS], int lim, int nexs)
void print_avgs(int avg], int nexs)
main()
{ int no_of_stds, no_of_exams
int avgCOLS]
int scoresMAX]COLS]
float wtsCOLS]
printf("***Weighted Average of Scores***\n\n")
no_of_exams = getwts(wts)
no_of_stds = read_scores(scores, MAX, no_of_exams)
wtd_avg(scores, no_of_stds, no_of_exams, wts)
print_scores(scores, no_of_stds, no_of_exams)
avg_scores(scores, avg, no_of_stds, no_of_exams)
print_avgs(avg, no_of_exams)
}
Figure 9.2: Driver for Student Scores Program
9.1. TWO DIMENSIONAL ARRAYS 345
/* File: wtdavg.c - continued */
/* Gets the number of exams and weights for the exams flushes
the input buffer and returns the number of exams.
*/
int getwts(float wts])
{ int i, n
printf("Number of exams: ")
scanf("%d", &n)
for (i = 1 i <= n i++) {
printf("Percent Weight for Exam %d: ", i)
scanf("%f", &wtsi])
}
while (getchar() != '\n')
return n
}
Figure 9.3: Code for getwts()
the input. The function returns the number of exams, n.
We will assume that the data for the student scores is stored in a le in the format of one line
per student, with each line containing the student id followed by the exam scores. To read this
data into a two dimensional array, we must rst open the input le. This is done by the function
openfile() shown in Figure 9.4, which prompts for the le name and tries to open the le. If
the le opens successfully, the le pointer is returned. Otherwise, the function prints a message
and asks the user to retype the le name. The user may quit at any time by typing a newline
or end of le. If an end of le is typed or the typed string is empty, the program is terminated.
Once the input le is opened, we read data items into the array, lling in the elements one row
(student) at a time. We use two index variables, row and col, varying the row to access each row
in turn
and, within each row, we vary col to access elements of each column in turn. We will
need a doubly nested loop to read the data in this manner. The function is given the number of
students, the variable stds, and the number of exams, nexs. We will use column 0 to store the
student id numbers and the next nexs columns to store the scores. Thus, in each row, we read
nexs+1 data values into the array. This is done by the function, read scores(), also shown in
Figure 9.4. The input le is rst opened using openfile(), and the data records are read into the
array called ex]] within the function. The function returns the number of records read either
when EOF is reached or when the array is lled. Each integer data item is read from a le, fp, into
a temporary variable, n. This value is then assigned to the appropriate element, exrow]col].
When all data has been read, the input le is closed and the number of records read is returned.
Notice in main() in Figure 9.2, we pass the 2D array to read scores() just as we did for
one dimensional arrays, passing the array name. As we shall see in the next section, the array
346 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: wtdavg.c - continued */
/* Opens the input file and returns the file pointer. */
FILE *openinfile(void)
{ FILE *fp
char infile25]
printf("Input File, RETURN to quit: ")
while (gets(infile)) {
if (!*infile) exit(0) /* empty string, exit */
fp = fopen(infile, "r")
if (!fp) { /* no such file, continue */
printf("Unable to open input file - retype\n")
continue
}
else return fp /* file opened, return fp */
}
exit(0) /* end of file typed, exit */
}
/* Opens the input file and reads scores for nexs exams returns
the number of individual student records.
*/
int read_scores(int ex]COLS], int stds, int nexs)
{ int row, col, n, x
FILE * fp
fp = openinfile()
for (row = 0 row < stds row++)
for (col = 0 col <= nexs col++) {
x = fscanf(fp, "%d", &n)
if (x == EOF) {
fclose(fp)
return row
}
exrow]col] = n
}
fclose(fp)
return row
}
Figure 9.4: Code for openfile() and read scores()
9.1. TWO DIMENSIONAL ARRAYS 347
name is a pointer that allows indirect access to the array elements. The two dimensional array
as as argument must be declared in the function denition as a formal parameter. In Figure
9.4, we have declared it as ex]COL] with two sets of square brackets to indicate that it points
to a two dimensional array. In our declaration, we must include the number of columns in the
array because this species the size of each row. Recall, the two dimensional array is an array of
rows. Once the compiler knows the size of a row in the array, it is able to correctly determine the
beginning of each row.
The next function called in main() computes the weighted average for each row. The weighted
average for one record is just the sum of each of the exam score times the actual weight of that
exam. If the scores are in the array, ex]], then the following code will compute a weighted
average for a single row, row:
wtdavg = 0.0
for (col = 1 col <= nexs col++)
wtdavg += exrow]col] * wtscol] / 100.0
We convert the percent weight to the actual weight multiply by the score, and accumulate it in
the sum, wtdavg yielding a oat value. The wtdavg will be stored in the integer array, ex]],
after rounding to a nearest integer. If we simply cast wtdavg to an integer, it will be truncated.
To round to the nearest integer, we add 0.5 to wtdavg and then cast it to integer:
exrow]nexs + 1] = (int) (0.5 + wtdavg)
The weighted average is stored into the column of the array after the last exam score. The entire
function is shown in Figure 9.5
Computing average of each of the exams and the weighted average is simple. We just sum
each of the columns and divide by the number of items in the column, and is also shown in Figure
9.5. For each exam and for the weighted average column, the scores are added and divided by
lim, the number of rows in the array, using oating point computation. The result is rounded to
the nearest integer and stored in the array, avg]. Figure 9.6 shows the nal two functions for
printing the results.
Running the program with data le,wtdin.dat as follows:
3 70 76
52 92 80
53 95 56
54 48 52
55 98 95
57 100 95
61 100 65
62 95 76
63 86 65
70 100 90
71 73 73
75 94 79
produces the following sample session:
***Weighted Average of Scores***
348 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: wtdavg.c - continued */
/* Computes the weighted average of the exam scores in ex]] for
lim individuals, nexs number of exams, and weights given by wts].
*/
void wtd_avg(int ex]COLS], int lim, int nexs, float wts])
{ int i, j
float wtdavg
for (i = 0 i < lim i++) {
wtdavg = 0.0
for (j = 1 j <= nexs j++)
wtdavg += exi]j] * wtsj] / 100.0
exi]nexs + 1] = (int) (wtdavg + 0.5)
}
}
/* Averages exam and weighted average scores. */
void avg_scores(int ex]COLS], int avg], int lim, int nexs)
{ int i, j
for (j = 1 j <= nexs + 1 j++) {
avgj] = 0
for (i = 0 i < lim i++)
avgj] += exi]j]
avgj] = (int) (0.5 + (float) avgj] / (float) lim)
}
}
Figure 9.5: Code for wtd avg() and avg scores()
9.1. TWO DIMENSIONAL ARRAYS 349
/* File: wtdavg.c - continued */
/* Prints the scores for exams and the weighted average. */
void print_scores(int ex]COLS], int lim, int nexs)
{ int i, j
printf("ID #\t")
for (j = 1 j <= nexs j++)
printf("Ex%d\t", j) /* print the headings */
printf("WtdAvg\n")
for (i = 0 i < lim i++) { /* print the scores and wtd avg */
for (j = 0 j <= nexs + 1 j++)
printf("%d\t", exi]j])
printf("\n")
}
}
/* Prints the averages of exams and the average of the weighted average
of exams.
*/
void print_avgs(int avg], int nexs)
{ int i
for (i = 1 i <= nexs i++)
printf("Average for Exam %d = %d\n", i, avgi])
printf("Average of the weighted average = %d\n", avgnexs + 1])
}
Figure 9.6: Code for print scores() and print avgs()
350 CHAPTER 9. TWO DIMENSIONAL ARRAYS
Number of exams: 2
Percent Weight for Exam 1: 50
Percent Weight for Exam 2: 50
Input File, RETURN to quit: wtdin.dat
ID # Ex1 Ex2 WtdAvg
3 70 76 73
52 92 80 86
53 95 56 76
54 48 52 50
55 98 95 97
57 100 95 98
61 100 65 83
62 95 76 86
63 86 65 76
70 100 90 95
71 73 73 73
75 94 79 87
Average for Exam 1 = 88
Average for Exam 2 = 75
Average of the weighted average = 82
In this program, we have assumed that the input le contains only the data to be read, i.e. the
student id numbers and exam scores. Our read scores() function is written with this assumption.
However, the input le might also contain some heading information such as the course name and
column headings in the rst few lines of the le. We can easily modify read scores() to discard
the rst few lines of headings.
As a second example of application of two dimensional arrays, consider our previous payroll
example. In this case, the data items in a pay data record are not all of the same data type. The
id numbers are integers, whereas all the other items are oat. Therefore, we must use an array of
integers to store the id numbers, and a two dimensional oat array to store the rest of the data
record. The algorithm is no dierent from the program we developed in Chapter 7 that computed
pay. The dierence is that now we use a two dimensional array for all oat payroll data instead
of several one dimensional arrays. The id numbers are still stored in a separate one dimensional
array. Since the data structures are now dierent, we must recode the functions to perform the
tasks of getting data, calculating pay, and printing results, but still using the same algorithms.
The program driver and the header les are shown in Figure 9.7. The program declares an
integer array for id numbers and a two dimensional oat array for the rest of the data record. The
successive columns in the two dimensional array store the hours worked, rate of pay, regular pay,
overtime pay, and total pay, respectively. We have dened macros for symbolic names for these
index values. As in the previous version, the program gets data, calculates pay, and prints data.
The dierence is in the data structures used. Functions to perform the actual tasks are shown
in Figure 9.8 and 9.9 and included in the same program source le. Each function uses a two
dimensional array, payrec]]. The row index species the data record for a single id, and the
column index species a data item in the record. The data record also contains the total pay. A
sample interaction with the program, pay2rec.c, is shown below.
9.1. TWO DIMENSIONAL ARRAYS 351
/* File: pay2rec.c
Program calculates and stores payroll data for a number of id's.
The program uses a one dimensional array for id's, and a two
dimensional array for the rest of the pay record. The first column
is hours, the second is rate, the third is regular pay, the fourth
is overtime pay, and the fifth is total pay.
*/
#include <stdio.h>
#define MAX 10
#define REG_LIMIT 40
#define OT_FACTOR 1.5
#define HRS 0
#define RATE 1
#define REG 2
#define OVER 3
#define TOT 4
int get2data(int id], float payrec]TOT + 1], int lim)
void calc2pay(float payrec]TOT + 1], int n)
void print2data(int id], float payrec]TOT + 1], int n)
main()
{ int n = 0, idMAX]
float payrecMAX]TOT + 1]
printf("***Payroll Program - Records in 2 D arrays***\n\n")
n = get2data(id, payrec, MAX)
calc2pay(payrec, n)
print2data(id, payrec, n)
}
Figure 9.7: Driver for Payroll Program Using 2D Arrays
352 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: pay2rec.c - continued */
/* Gets id's in one array, and the rest of input data records
in a two dimensional array.
*/
int get2data(int id], float payrec]TOT + 1], int lim)
{ int n = 0
float x
while (n < lim) {
printf("ID <zero to quit>: ")
scanf("%d", &idn])
if (idn] <= 0)
return n
printf("Hours Worked: ")
scanf("%f", &x)
payrecn]HRS] = x
printf("Rate of Pay: ")
scanf("%f", &x)
payrecn]RATE] = x
n++
}
if (n == lim) {
printf("Table full, processing data\n")
return n
}
}
Figure 9.8: Code for Payroll Program Functions | get2data()
9.1. TWO DIMENSIONAL ARRAYS 353
/* Calculates pay for each id record in a two dimensional array. */
void calc2pay(float payrec]TOT + 1], int n)
{ int i
for (i = 0 i < n i++) {
if (payreci]HRS] <= REG_LIMIT) {
payreci]REG] = payreci]HRS] * payreci]RATE]
payreci]OVER] = 0
}
else {
payreci]REG] = REG_LIMIT * payreci]RATE]
payreci]OVER] = (payreci]HRS] - REG_LIMIT) *
OT_FACTOR * payreci]RATE]
}
payreci]TOT] = payreci]REG] + payreci]OVER]
}
}
/* Prints a table of payroll data for all id's. Id's in one array,
and the rest of the records in a two dim. array.
*/
void print2data(int id], float payrec]TOT + 1], int n)
{ int i, j
printf("***PAYROLL: FINAL REPORT***\n\n")
printf("%10s%10s%10s%10s%10s%10s\n", "ID", "HRS",
"RATE", "REG", "OVER", "TOT")
for (i = 0 i < n i++) {
printf("%10d", idi])
for (j = 0 j <= TOT j++)
printf("%10.2f", payreci]j])
printf("\n")
}
}
Figure 9.9: Code for Payroll Program Functions | calc2pay() and print2data()
354 CHAPTER 9. TWO DIMENSIONAL ARRAYS
Sample Session:
***Payroll Program - Records in 2 D arrays***
ID <zero to quit>: 5
Hours Worked: 30
Rate of Pay: 10
ID <zero to quit>: 10
Hours Worked: 50
Rate of Pay: 12
ID <zero to quit>: 0
***PAYROLL: FINAL REPORT***
ID HRS RATE REG OVER TOT
5 30.00 10.00 300.00 0.00 300.00
10 50.00 12.00 480.00 180.00 660.00
9.2 Implementing Multi-Dimensional Arrays
In the last section we saw how we can use two dimensional arrays | how to declare them, pass
them to functions, and access the date elements they contain using array indexing notation. As
with one dimensional arrays, we can access elements in a 2D array using pointers as well. In
order to understand how this is done, in this section we look at how multi dimensional arrays are
implemented in C.
As we saw in Chapter 7, a one dimensional array is stored in a set of contiguous memory cells
and the name of the array is associated with a pointer cell to this block of memory. In C, a two
dimensional array is considered to be a one dimensional array of rows, which are, themselves,
one dimensional arrays. Therefore, a two dimensional array of integers, AA]], is stored as a
contiguous sequence of elements, each of which is a one dimensional array. The rows are stored
in sequence, starting with row 0 and continuing until the last row is stored, i.e. AA0] is stored
rst, then AA1], then AA2], and so on to AAMAX-1]. Each of these \elements" is an array, so
is stored as a contiguous block of integer cells as seen in Figure 9.10. This storage organization
for two dimensional arrays is called row major order. The same is true for higher dimensional
arrays. An n dimensional array is considered to be a one dimensional array whose elements are,
themselves, arrays of dimension n ; 1. As such, in C, an array of any dimension is stored in row
major order in memory.
With this storage organization in mind, let us look at what implications this has to referencing
the array with pointers. Recall that an array name (without an index) represents a pointer to the
rst object of the array. So the name, AA, is a pointer to the element AA0]. iBut, AA0] is a one
dimensional array
so, AA0] points to the rst object in row 0, i.e. AA0] points to AA0]0].
Similarly, for any k AAk] points to the beginning of the kth row, i.e. AAk] is the address of
AAk]0]. Since AAk] points to AAk]0], *AAk] accesses AAk]0], an object of type integer.
If we add 1 to the pointer AAk], the resulting pointer will point to the next integer type element,
i.e. the value of AAk]1]. In general, AAk] + j points to AAk]j], and *(AAk] + j) accesses
the value of AAk]j]. This is shown in Tables 9.1 and 9.2. Each AAk] points to an integer type
object. When an integer is added to the pointer AAk], the resulting pointer points to the next
object of the integer type.
9.2. IMPLEMENTING MULTI-DIMENSIONAL ARRAYS 355
AA0]
AA1]
...
AAMAX-1]
Figure 9.10: A Two Dimensional Array in Row Major Order
AA 0] &AA 0] 0]
AA 1] &AA 1] 0]
AA 2] &AA 2] 0]
AA k] &AA k] 0]
AA 0] + 1 &AA 0] 1]
AA 0] + j &AA 0] j]
AA k] + j &AA k] j]
Table 9.1: Array Pointers and Sub-Arrays
* AA 0] AA 0] 0]
AA k] AA k] 0]
(AA 0] + 1) AA 0] 1]
(AA 0] + j) AA 0] j]
(AA k] + j) AA k] j]
Table 9.2: Dereferencing Array Pointers
356 CHAPTER 9. TWO DIMENSIONAL ARRAYS
* AA AA 0] &AA 0] 0]
AA + 1 AA 0] + 1 &AA 0] 1]
AA + j AA 0] + j &AA 0] j]
(AA + 1) AA 1] &AA 1] 0]
(AA + k) AA k] &AA k] 0]
(* AA) *AA 0] AA 0] 0]
(* (AA + 1)) *AA 1] AA 1] 0]
(* (AA + k) + j) *(AA k] + j) AA k] j]
Table 9.3: Pointer Equivalence for Two Dimensional Arrays
The name, AA, is the name of the entire array, whose elements are themselves arrays of integers.
Therefore, AA points to the rst object in this array of arrays, i.e. AA points to the array AA0]. The
addresses represented by AA and AA0] are the same
however, they point to objects of dierent
types. AA0] points to AA0]0], so it is an integer pointer. AA points to AA0], so it is a pointer
to an integer pointer. If we add 1 to AA, the resulting pointer, AA + 1, points to the array AA1],
and AA + k points to the array AAk]. When we add to a pointer to some type, we point to the
next object of that type. Therefore, adding to AA and AA0] result in pointers to dierent objects.
Adding 1 to AA results in a pointer that points to the next array or row, i.e. AA1]
whereas adding
1 to AA0] results in a pointer that points to AA0]1]. Dereferencing such a pointer, *(AA + k),
accesses AAk], which as we saw, was &AAk]0]. It follows that *(*(AA + k) + j) accesses the
integer, AAk]j]. This pointer equivalence for two dimensional arrays is shown in Table 9.3.
The C compiler converts array indexing to indirect access by dereferenced pointers as shown
in the table
thus, all array access is indirect access When we pass a 2D array to a function, we
pass its name (AA). The function can access elements of the array argument either by indexing
or by pointers. We generally think of a two dimensional array as a table consisting of rows and
columns as seen in Figure 9.11. As such, it is usually easiest to access the elements by indexing
however, the pointer references are equally valid as seen in the gure.
The relationships between dierent pointers for a two dimensional array is further illustrated
with the program shown in Figure 9.12. The two-dimensional array, a, is not an integer pointer,
it points to the array of integers, a0]. However, *a is an integer pointer
it points to an integer
object, a0]0]. To emphasize this point, we initialize an integer pointer, intptr to *a, i.e. a0].
The initial value of intptr is the address of a0]0]. We next print the values of a and *a, which
are the same address even though they point to dierent types of objects. In the for loop, we
print the value of a + i, which is the same as that of ai] even though they point to dierent
types of objects. In the inner for loop, we print the address of the ith row and the j th column
element of the row major array using pointers:
*a + COL * i + j
The same value is printed using the address of operator, &ai]j]. Finally, the value of ai]j]
is printed using array indices as well as by dereferencing pointers, i.e. *(*(a + i) + j).
The value of intptr, initialized to *a, is incremented after each element value is printed
9.2. IMPLEMENTING MULTI-DIMENSIONAL ARRAYS 357
AA0]+1 AA0]+2 AA0]+MAXCOL-1
? ? ?
AA - AA0] -
AA + 1 - AA1] -
AA + 2 - AA2] -
..
AA + MAX - 1 -AAMAX-1] -
Figure 9.11: Pointers and Two Dimensional Arrays
making it point to the next element. The value of intptr is printed as it is incremented. Observe
that it prints the address of each element of the array in one row, and proceeds to the next row
in sequence. This shows that arrays are stored in row major form.
Finally, the function, print2aray() is used to print the two dimensional array in rows and
columns. The output of a sample run is shown below.
***2D Arrays, Pointers ***
array (row) pointer a = 65474, *a = 65474
a + 0 = 65474
*a + COL * 0 + 0 = 65474 intptr = 65474
&a0]0] = 65474
a0]0] = 12 *(*(a + 0) + 0) = 12
*a + COL * 0 + 1 = 65476 intptr = 65476
&a0]1] = 65476
a0]1] = 24 *(*(a + 0) + 1) = 24
*a + COL * 0 + 2 = 65478 intptr = 65478
&a0]2] = 65478
a0]2] = 29 *(*(a + 0) + 2) = 29
a + 1 = 65480
*a + COL * 1 + 0 = 65480 intptr = 65480
&a1]0] = 65480
a1]0] = 23 *(*(a + 1) + 0) = 23
358 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: ar2ptr.c
Other Source Files: ar2util.c
Header Files: ar2def.h, ar2util.h
Program shows relations between arrays and pointers for 2 dimensional
arrays.
*/
#include <stdio.h>
#define ROW 2
#define COL 3
print2aray(int x]COL], int r, int c)
main()
{ int i, j, *intptr, aROW]COL] =
{ {12, 24, 29}, {23, 57, 19} }
printf("***2D Arrays, Pointers ***\n\n")
intptr = *a
printf("array (row) pointer a = %u, *a = %u\n", a, *a)
for (i = 0 i < ROW i++) {
printf("a + %d = %u\n", i, a + i)
for (j = 0 j < COL j++) {
printf("*a + COL * %d + %d = %u intptr = %u\n",
i, j, *a + COL * i + j, intptr)
printf("&a%d]%d] = %u\n", i, j, &ai]j])
printf("a%d]%d] = %d *(*(a + %d) + %d) = %d\n",
i, j, ai]j],
i, j, *(*(a + i) + j))
intptr++
}
}
print2aray(a, ROW, COL)
}
/* This Function prints a two dimensional integer array. */
print2aray(int x]COL], int r, int c)
{ int i, j
printf("\nThe two dimensional array is:\n\n")
for (i = 0 i < r i++) {
for (j = 0 j < c j++)
printf("%d ", xi]j])
printf("\n")
}
}
Figure 9.12: Program Illustrating 2D Array Pointers
9.3. ARRAYS OF STRINGS 359
*a + COL * 1 + 1 = 65482 intptr = 65482
&a1]1] = 65482
a1]1] = 57 *(*(a + 1) + 1) = 57
*a + COL * 1 + 2 = 65484 intptr = 65484
&a1]2] = 65484
a1]2] = 19 *(*(a + 1) + 2) = 19
The two dimensional array is:
12 24 29
23 57 19
As we mentioned in the last section, when a two dimensional array is passed to a function,
the parameter declaration in the function must include the number of columns. We can now see
why this is so. The number of columns in a row species the size of each row in the array of rows.
Since the passed parameter is a pointer to a row object, it can be incremented and dereferenced,
as shown in Table 9.3, to access the elements of the two dimensional array. The compiler must
know the size of the row in order to be able to increment the pointer to the next row.
As we stated earlier, multi-dimensional arrays are arrays of arrays just like two dimensional
arrays. An n dimensional array is an array of n ; 1 dimensional arrays. The same general approach
applies as for two dimensional arrays. When passing an n dimensional array, the declaration of
the formal parameter must specify all index ranges except for the rst index.
As was seen in the program in Figure 9.12, multi-dimensional arrays may also be initialized in
declarations by specifying constant initializers within braces. Each initializer must be appropriate
for the corresponding lower dimensional array. For example, a two dimensional array may be
initialized as follows:
int x2]3] = { {10, 23}, {0, 12, 17} }
The array has two elements, each of which is an array of three elements. The rst initializer
initializes the rst row of x. Since only the rst two elements of the row are specied, the third
element is zero. The second element initializes the second row. Thus, x is initialized to the array:
10 23 0
0 12 17
9.3 Arrays of Strings
Besides data base applications, another common application of two dimensional arrays is to store
an array of strings. In this section we see how an array of strings can be declared and operations
such as reading, printing and sorting can be performed on them.
A string is an array of characters
so, an array of strings is an array of arrays of characters.
Of course, the maximum size is the same for all the strings stored in a two dimensional array. We
can declare a two dimensional character array of MAX strings of size SIZE as follows:
char namesMAX]SIZE]
360 CHAPTER 9. TWO DIMENSIONAL ARRAYS
names0] - J o h n n0
names1] - S u e n0
names2] -
...
namesMAX-1] -
Figure 9.13: An Array of Strings
Since names is an array of character arrays, namesi] is the ith character array, i.e. it points
to the ith character array or string, and may be used as a string of maximum size SIZE - 1. As
usual with strings, a NULL character must terminate each character string in the array. We can
think of an array of strings as a table of strings, where each row of the table is a string as seen in
Figure 9.13.
We will need an array of strings in our next task to read strings, store them in an array, and
print them.
NAMES: Read and store a set of strings. Print the strings.
We can store a string into namesi] by reading a string using gets() or by copying one into
it using strcpy(). Since our task is to read strings, we will use gets(). The algorithm is simple:
while array not exhausted and not end of file,
read a string into an array element
print out the strings in the array of strings
We will organize the program in several source les since we will be using some of the functions
in several example programs. The program driver and the header le are shown in Figure 9.14.
The program reads character strings into an array, in this case, names. The program can, of
course, serve to read in any strings. The for loop in main() reads strings into an array using
gets() to read a string into namesn], the nth row of the array. That is, the string is stored
where namesn] points to. The variable n keeps track of the number of names read. The loop is
terminated either if the number of names equals MAX, or when gets() returns NULL indicating end
of le has been reached. Next, the program calls on printstrtab() to print the names stored in
the two dimensional array, names. The arguments passed are the array of strings and the number
of strings, n.
The function, printstrtab() is included in the le strtab.c and its prototype is included
in the le strtab.h. Remember, the second range of the two dimensional array of strings must
be specied in the formal parameter denition, otherwise the number of columns in a row are
unknown and the function cannot access successive rows correctly. A sample interaction for the
compiled and linked program is shown below:
Sample Session:
***Table of Strings - Names***
9.3. ARRAYS OF STRINGS 361
/* File: strtab.h */
#define SIZE 31 /* maximum size of a name plus a NULL */
void printstrtab(char strtab]SIZE], int n)
/* File: names.c
Other Source Files: strtab.c
Header Files: strtab.h
This program reads a set of names or strings into a two
dimensional array. It then prints out the names.
*/
#include <stdio.h>
#define MAX 10
#include "strtab.h"
main()
{ int n /* number of names */
char namesMAX]SIZE] /* 2-d array of names */
printf("***Table of Strings - Names***\n\n")
printf("Enter one name per line, EOF to terminate\n")
for (n = 0 (n < MAX) && gets(namesn]) n++)
if (n == MAX)
printf("\n**Table full - input terminated\n")
printstrtab(names, n)
}
/* File: strtab.c */
#include <stdio.h>
#include "strtab.h"
/* Prints n strings in the array strtab]]. */
void printstrtab(char strtab]SIZE], int n)
{ int k
printf("Names are:\n")
for (k = 0 k < n k++)
puts(strtabk])
}
Figure 9.14: Code for String Table Driver
362 CHAPTER 9. TWO DIMENSIONAL ARRAYS
Enter one name per line, EOF to terminate
John Smith
David Jones
Helen Peterson
Maria Schell
^D
Names are:
John Smith
David Jones
Helen Peterson
Maria Schell
9.3.1 String Sorting and Searching
Our next couple of tasks are simple and build on the last one. In one task we search (sequentially)
for a string and in another we sort a set of strings.
SRCHSTR: Search for a key string in a set of strings.
We will use a function, srchstr(), to search for a string in an array of a specied size. The
function returns either a valid index where the string is found or it returns -1 to indicate failure.
The algorithm is simple enough and the implementation of a test program driver is shown in
Figure 9.15.
The le strtab.h includes the prototypes for functions in le strtab.c. Observe the ini-
tialization of the two dimensional array names]] using constant initializers written in braces
separated by commas. Each initializer initializes a one dimensional string array written as a string
constant. The program calls srchstrtab() searching for the string "John Smith", and prints the
returned index value.
As we have seen in Chapter 8, the library function strcmp() is used to compare two strings.
The function returns zero if the argument strings are equal. A sequential search process for strings
is easily developed by modifying the sequential search function of Chapter 10 replacing the equality
operator with the function strcmp() to compare two strings.
In the function, srchstrtab(), we compare each string in the array with the desired string
until we either nd a match or the array is exhausted. The function call requires the name of the
array of strings, the number of valid elements in the array, and the item to be searched for. For
example, suppose we wish to search for a string, key, in the array, names with n string elements,
then, the function call would be:
k = srchstrtab(names, n, key)
The value returned is assigned to an integer variable, k. If successful, srchstrtab() returns the
index where the string was found
otherwise, it returns -1. The function is shown in Figure 9.16. In
the for loop, the string that strtabi] points to is compared with the string that key points to. If
they are equal, strcmp() returns zero and the value of i is returned by the function. Otherwise, i
is incremented, and the process is repeated. The loop continues until the valid array is exhausted,
in which case -1 is returned. Again, the formal parameter denition for the two dimensional array,
x, requires the size of the second dimension, SIZE. A sample run of the program is shown below:
9.3. ARRAYS OF STRINGS 363
/* File: strsrch.c
Other Source Files: strtab.c
Header Files: strtab.h
This program searches for a string (key) in a set of strings
in a two dimensional array. It prints the index where key is found.
It prints -1, if the string is not found.
*/
#include <stdio.h>
#define MAX 10
#include "strtab.h"
main()
{ int k
char namesMAX]SIZE] = { "John Jones", "Sheila Smith",
"John Smith", "Helen Kent"}
printf("***String Search for John Smith***\n\n")
k = srchstrtab(names, 4, "John Smith")
printf("John Smith found at index %d\n", k)
}
Figure 9.15: Driver for String Search Program
/* File: strtab.c - continued */
#include <string.h>
/* Searches a string table strtab]] of size n for a string key. */
int srchstrtab(char strtab]SIZE], int n, char key])
{ int i
for (i = 0 i < n i++)
if (strcmp(strtabi], key) == 0)
return i
return -1
}
Figure 9.16: Code for srchstrtab()
364 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: strsort.c
Other Source Files: strtab.c
Header Files: strtab.h
This program sorts a set of strings in a two dimensional array.
It prints the unsorted and the sorted set of strings.
*/
#include <stdio.h>
#define MAX 10
#include "strtab.h"
main()
{ int n
char namesMAX]SIZE] = { "John Jones", "Sheila Smith",
"John Smith", "Helen Kent"}
printf("***String Array - unsorted and sorted***\n\n")
printf("Unsorted ")
printstrtab(names, 4)
sortstrtab(names, 4)
printf("Sorted ")
printstrtab(names, 4)
}
Figure 9.17: Driver for Sorting Strings Program
***String Search for John Smith***
John Smith found at index 2
Our next task calls for sorting a set of strings.
SORTSTR: Sort a set of strings. Print strings in unsorted and in sorted order.
The algorithm is again very simple and we implement it in the program driver. The driver
simply calls on sortstrtab() to sort the strings and prints the strings, rst unsorted and then
sorted. A prototype for sortstrtab() is included in le strtab.h and the driver is shown in
Figure 9.17. An array of strings is initialized in the declaration and the unsorted array is printed.
Then, the array is sorted, and the sorted array is printed.
Sorting of an array of strings is equally straight forward. Let us assume, the array of strings is
to be sorted in increasing ASCII order, i.e. a is less than b, b is less than c, A is less than a, and so
on. We will use the selection sort algorithm from Chapter 10. Two nested loops are needed
the
inner loop moves the largest string in an array of some eective size to the highest index in the
array, and the outer loop repeats the process with a decremented eective size until the eective
size is one. The function is included in le strtab.c and shown in Figure 9.18. The function
9.3. ARRAYS OF STRINGS 365
/* File: strtab.c - continued */
/* Sorts an array of strings. The number of strings in the
array is lim.
*/
void sortstrtab(char strtab]SIZE], int lim)
{ int i, eff_size, maxpos = 0
char tmpSIZE]
for (eff_size = lim eff_size > 1 eff_size--) {
for (i = 0 i < eff_size i++)
if (strcmp(strtabi], strtabmaxpos]) > 0)
maxpos = i
strcpy(tmp, strtabmaxpos])
strcpy(strtabmaxpos], strtabeff_size-1])
strcpy(strtabeff_size - 1], tmp)
}
}
Figure 9.18: Code for sortstrtab()
is similar to the numeric selection sort function, except that we now use strcmp() to compare
strings and strcpy() to swap strings. A sample session is shown below:
***String Array - unsorted and sorted***
Unsorted Names are:
John Jones
Sheila Smith
John Smith
Helen Kent
Sorted Names are:
Helen Kent
John Jones
John Smith
Sheila Smith
In our example program, the entire strings are compared. If we wish to sort by last name, we
could modify our function to nd and compare only the last names.
366 CHAPTER 9. TWO DIMENSIONAL ARRAYS
9.4 Arrays of Pointers
As seen in the last example, sorting an array of strings requires swapping the strings which can
require copying a lot of data. For eciency, it is better to avoid actual swapping of data whenever
a data item is large, such as a string or an entire data base record. In addition, arrays may be
needed in more than one order
for example, we may need an exam scores array sorted by Id
number and by weighted scores
or, we may need strings in both an unsorted form and a sorted
form. In either of these cases, we must either keep two copies of the data, each sorted dierently,
or nd a more ecient way to store the data structure. The solution is to use pointers to elements
of the array and swap pointers. Consider some examples:
int data1, data2, *ptr1, *ptr2, *save
data1 = 100 data2 = 200
ptr1 = &data1 ptr2 = &data2
We could swap the values of the data and store the swapped values in data1 and data2 or we
could simply swap the values of the pointers:
save = ptr1
ptr1 = ptr2
ptr2 = save
We have not changed the values in data1 and data2
but ptr1 now accesses data2 and ptr2
access data1. We have swapped the pointer values so they point to objects in a dierent order.
We can apply the same idea to strings:
char name1] = "John"
char name2] = "Dave"
char *p1, *p2, *save
p1 = name1
p2 = name2
Pointers p1 and p2 point to strings name1 and name2. We can now swap the pointer values so p1
and p2 point to name2 and name1, respectively.
In general, an array of pointers can be used to point to an array of data items with each
element of the pointer array pointing to an element of the data array. Data items can be accessed
either directly in the data array, or indirectly by dereferencing the elements of the pointer array.
The advantage of a pointer array is that the pointers can be reordered in any manner without
moving the data items. For example, the pointer array can be reordered so that the successive
elements of the pointer array point to data items in sorted order without moving the data items.
Reordering pointers is relatively fast compared to reordering large data items such as data records
or strings. This approach saves a lot of time, with the additional advantage that the data items
remain available in the original order. Let us see how we might implement such a scheme.
STRPTRS: Given an array of strings, use pointers to order the strings in sorted form, leaving
the array unchanged.
We will use an array of character pointers to point to the strings declared as follows:
9.4. ARRAYS OF POINTERS 367
0 - j o h n n0
1 - y u r i n0
2 - s u e n0
3 - c a r o l n0
4 - d a v i d n0
..
MAX-1
captionUnsorted Pointers to Strings
char * nameptrMAX]
The array, nameptr], is an array of size MAX, and each element of the array is a character pointer.
It is then possible to assign character pointer values to the elements of the array
for example:
nameptri] = "John Smith"
The string "John Smith" is placed somewhere in memory by the compiler and the pointer to the
string constant is then assigned to nameptri]. It is also possible to assign the value of any string
pointer to nameptri]
for example, if s is a string, then it is possible to assign the pointer value
s to nameptri]:
nameptri] = s
In particular, we can read strings into a two dimensional array, names]], and assign each string
pointer, namesi] to the ith element of the pointer array, nameptr]:
for (i = 0 i < MAX && gets(namesi]) i++)
nameptri] = namesi]
The strings can then be accessed either by namesi] or by nameptri] as seen in Figure 9.4. We
can then reorder the pointers in nameptr] so that they successively point to the strings in sorted
order as seen in Figure 9.4. We can then print the strings in the original order by accessing them
through namesi] and print the strings in sorted order by accessing them through nameptri].
Here is the algorithm:
while not end of file and array not exhausted,
read a string
store it in an array of strings and
assign the string to an element of a pointer array
access the array of strings and print them out
access the array of pointers and print strings that point to
368 CHAPTER 9. TWO DIMENSIONAL ARRAYS
0 - j o h n n0
1 - y u r i n0
2 - s u e n0
3 - c a r o l n0
4 - d a v i d n0
..
MAX-1
captionSorted Pointers to Strings
The program driver, including a prototype for sortptrs() is shown in Figure 9.19. It declares
a two dimensional array of strings, names]], and an array of character pointers, nameptr].
It then reads strings into names]], and assigns each string pointer namesi] to nameptri].
The function sortptrs() is then called to reorder nameptr] so the successive pointers of the
array point to the strings in sorted order. Finally, strings are printed in original unsorted order
by accessing them through namesi] and in sorted order by accessing them through nameptri].
The function sortptrs() uses the selection sort algorithm modied to access data items
through pointers. It repeatedly moves the pointer to the largest string to the highest index of an
eective array. The implementation of sorting using pointers to strings is shown in Figure 9.20.
The algorithm determines maxpos, the index of the pointer to the largest string. The pointer at
maxpos is then moved to the highest index in the eective array. The array size is then reduced,
etc.
Sample Session:
***Arrays of Pointers - Sorting by Pointers***
Enter one name per line, EOF to terminate
john
yuri
sue
carol
david
^D
The unsorted names are:
john
yuri
sue
carol
david
9.4. ARRAYS OF POINTERS 369
/* File: ptraray.c
This program uses an array of pointers. Elements of the array
point to strings. The pointers are reordered so that they
point to the strings in sorted order. Unsorted and sorted
strings are printed out.
*/
#include <stdio.h>
#define TRUE 1
#define FALSE 0
#define MAX 10 /* max number of names */
#define SIZE 31 /* size of names plus one for NULL */
void sortptrs(char * nameptr], int n)
main()
{ int i /* counter */
int n /* number of names read */
char namesMAX]SIZE] /* 2-d array of names */
char *nameptrMAX] /* array of ptrs - used to point to names */
printf("***Arrays of Pointers - Sorting by Pointers***\n\n")
/* read the names into the 2-d array */
printf("Enter one name per line, ")
printf("EOF to terminate\n")
for (n = 0 gets(namesn]) && n < MAX n++)
nameptrn] = namesn] /* assign string pointer */
/* to a char pointer in the */
/* array of pointers. */
if (n == MAX)
printf("\n***Only %d names allowed***\n", MAX)
printf("The unsorted names are:\n")
/* print the names */
for (i = 0 i < n i++)
puts(namesi]) /* access names in stored array.*/
sortptrs(nameptr, n) /* sort pointers */
printf("The sorted names are:\n")
for (i = 0 i < n i++) /* print sorted names, */
puts(nameptri]) /* accessed via array of pointers. */
}
Figure 9.19: Driver for Sorting Pointer Array Program
370 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: ptraray.c - continued */
/* The elements of the array of pointers nameptr] point to
strings. The pointer array is reordered so the pointers
point to strings in sorted order. The function uses selection
sort algorithm.
*/
void sortptrs(char * nameptr], int n)
{ int i, eff_size, maxpos = 0
char *tmpptr
for (eff_size = n eff_size > 1 eff_size--) {
for (i = 0 i < eff_size i++)
if (strcmp(nameptri],nameptrmaxpos]) > 0)
maxpos = i
tmpptr = nameptrmaxpos]
nameptrmaxpos] = nameptreff_size-1]
nameptreff_size-1] = tmpptr
}
}
Figure 9.20: Code for sortptrs()
9.5. AN EXAMPLE: LINEAR ALGEBRAIC EQUATIONS 371
The sorted names are:
carol
david
john
sue
yuri
Reordering of pointers, so they point to data items in sorted order, is referred to as sorting by
pointers. When the data items are large, such as data records or strings, this is the preferred way
of sorting because it is far more ecient to move pointers than it is to move entire data records.
9.5 An Example: Linear Algebraic Equations
As our nal example program using two dimensional arrays in this chapter, we develop a program
to solve systems of simultaneous linear equations. A set of linear algebraic equations, also called
simultaneous equations, occur in a variety of mathematical applications in science, engineering,
economics, and social sciences. Examples include: electronic circuit analysis, econometric analysis,
structural analysis, etc. In the most general case, the number of equations, n, may be dierent
from the number of unknowns, m
thus, it may not be possible to nd a unique solution. However,
if n equals m, there is a good chance of nding a unique solution for the unknowns.
Our next task is to solve a set of linear algebraic equations, assuming that the number of
equations equals the number of unknowns:
LINEQNS: Read the coecients and the right hand side values for a set of linear equations
solve the equations for the unknowns.
The solution of a set of linear equations is fairly complex. We will rst review the process of
solution and then develop the algorithm in small parts. As we develop parts of the algorithm,
we will implement these parts as functions. The driver will just read the coecients, call on a
function to solve the equations, and call a function to print the solution.
Let us start with an example of a set of three simultaneous equations in three unknowns: x1,
x2, and x3.
1 x1 + 2 x2 + 3 x3 = 6
2 x1 + 3 x2 + 1 x3 = 6
1 x1 + 0 x2 + 2 x3 = 3
We can use arrays to represent such a set of equations
a two dimensional array to store the
coecients, a one dimensional array to store the values of the unknowns when solved, and another
one dimensional array to store the values on the right hand side. Later, we will include the right
hand side values as an additional column in the two dimensional array of coecients. Each row
of the two dimensional array stores the coecients of one of the equations. Since the array index
in C starts at 0, we will assume the unknowns are the elements x0], x1], and x2]. Similarly,
the elements in row zero are the coecients in the equation number 0, the elements in row one
are for equation number one, and so forth.
Then using arrays, a general set of n linear algebraic equations with m unknowns may be
expressed as shown below:
372 CHAPTER 9. TWO DIMENSIONAL ARRAYS
a0]0] * x0] + a0]1]*x1] + ...+ a0]m - 1] * xm - 1] = y0]
a1]0] * x0] + a1]1]*x1] + ...+ a1]m - 1] * xm - 1] = y1]
...
an-1]0]*x0] + ... + an-1]m - 1]*xm - 1] = yn-1]
The unknowns and the right hand side are assumed to be elements of one dimensional arrays:
x0], x1],: : :, xm - 1] and y0], y1],: : :, yn - 1], respectively. The coecients are
assumed to be elements of a two dimensional array: ai]j] for i = 0 : : : n ; 1 and j =
0 : : : m ; 1. The coecients of each equation correspond to a row of the array. For our discussion
in this section, we assume that m equals n. With this assumption, it is possible to nd a unique
solution of these equations unless the equations are linearly dependent, i.e. some equations are
linear combinations of others.
A common method for solving such equations is called the Gaussian elimination method.
The method eliminates (i.e. makes zero) all coecients below the main diagonal of the two
dimensional array. It does so by adding multiples of some equations to others in a systematic way.
The elimination makes the array of new coecients have an upper triangular form since the lower
triangular coecients are all zero.
The modied equivalent set of n equations in m = n unknowns in the upper triangular form
have the appearance shown below:
a0]0]*x0]+ a0]1]*x1] + ... + a0]n-1] *xn-1] = y0]
a1]1]*x1] + ... + a1]n-1] *xn-1] = y1]
a2]2]*x2]..+ a2]n-1] *xn-1] = y2]
an-1]n-1]*xn-1]= yn-1]
The upper triangular equations can be solved by back substitution. Back substitution rst solves
the last equation which has only one unknown, xn-1]. It is easily solved for this value | xn-1]
= yn-1]/an-1]n-1]. The next to the last equation may then be solved | since xn-1] has
been determined already, this value is substituted in the equation, and this equation has again only
one unknown, xn-2]. The unknown, xn-2], is solved for, and the process continues backward
to the next higher equation. At each stage, the values of the unknowns solved for in the previous
equations are substituted in the new equation leaving only one unknown. In this manner, each
equation has only one unknown which is easily solved for.
Let us take a simple example to see how the process works. For the equations:
1 * x0] + 2 * x1] + 3 * x2] = 6
2 * x0] + 3 * x1] + 1 * x2] = 6
1 * x0] + 0 * x1] + 2 * x2] = 3
We rst reduce to zero the coecients in the rst column below the main diagonal (i.e. array index
zero). If the rst equation is multiplied by -2 and added to the second equation, the coecient in
the second row and rst column will be zero:
1 * x0] + 2 * x1] + 3 * x2] = 6
0 * x0] - 1 * x1] - 5 * x2] = -6
1 * x0] + 0 * x1] + 2 * x2] = 3
Similarly, if the rst equation is multiplied by -1 and added to the third equation, the coecient
in the third row and rst column will be zero:
9.5. AN EXAMPLE: LINEAR ALGEBRAIC EQUATIONS 373
1 * x0] + 2 * x1] + 3 * x2] = 6
0 * x0] - 1 * x1] - 5 * x2] = -6
0 * x0] - 2 * x1] - 1 * x2] = -3
Coecients in the rst column below the main diagonal are now all zero, so we do the same for
the second column. In this case, the second equation is multiplied by a multiplier and added to
equations below the second
thus, multiplying the second equation by -2 and adding to the third
makes the coecient in the second column zero:
1 * x0] + 2 * x1] + 3 * x2] = 6
0 * x0] - 1 * x1] - 5 * x2] = -6
0 * x0] + 0 * x1] + 9 * x2] = 9
We now have equivalent equations with an upper triangular form for the non-zero coecients.
The equations can be solved backwards | the last equation gives us x2] = 1. Substituting the
value of x2] in the next to the last equation and solving for x1] gives us x1] = 1. Finally,
substituting x2] and x1] in the rst equation gives us x0] = 1.
From the above discussion, we can see that a general algorithm involves two steps: modify
the coecients of the equations to an upper triangular form, and solve the equations by back
substitution.
Let us rst consider the process of modifying the equations to an upper triangular form. Since
only the coecients and the right hand side values are involved in the computations that modify
the equations to upper triangular form, we can work with these items stored in an array with n
rows and n + 1 columns (the extra column contains the right hand side values).
Let us assume the process has already reduced to zero the rst k ; 1 columns below the main
diagonal, storing the modied new values of the elements in the same elements of the array. Now,
it is time to reduce the kth lower column to zero (by lower column, we mean the part of the column
below the main diagonal). The situation is shown in below:
a0]0] a0]1] ... a0]k] ... a0]n]
0 a1]1] ... a1]k] ... a1]n]
0 0 a2]2]... a2]k] ... a2]n]
... ... ... ... ... ...
0 0 0... ak]k] ak]k+1].. ak]n]
0 0 0... ak+1]k] ... ak+1]n]
0 0 0... ak+2]k] ... ak+2]n]
... ... ... ... ... ...
0 0 0... an-1]k] ... an-1]n]
The nth column represents the right hand side values with ai]n] equal to yi]. We multiply the
kth row by an appropriate multiplier and add it to each row with index greater than k. Assuming
that ak]k] is non-zero, the kth row multiplier for addition to the ith row (i > k) is:
-ai]k] / ak]k]
The kth row multiplied by the above multiplier and added to the ith row will make the new ai]k]
zero. The following loop will reduce to zero the lower kth column:
374 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* Algorithm: process_column
Reduces lower column k to zero.
*/
for (i = k + 1 i < n i++) { /* process rows k+1 to n-1 */
m = - ai]k] / ak]k] /* multiplier for ith row */
for (j = k j <= n j++) /* 0 thru k-1 cols. are zero */
ai]j] += m * ak]j] /* add kth row times m */
/* to ith row. */
}
However, before we can use the above loop to reduce the lower kth column to zero, we must
make sure that ak]k] is non-zero. If the current ak]k] is zero, all we need to do is exchange
this kth row with any higher indexed row with a non-zero element in the kth column. After the
exchange of the two rows, the new ak]k] will be non-zero. The above loop is then used to
reduce the lower kth column to zero. The non-zero element, ak]k] used in the multiplier is
called a pivot.
So, there are two steps involved in modifying the equations to upper triangular form: for each
row nd a pivot, and reduce the corresponding lower column to zero. If a non-zero pivot element is
not found, then one or more equations are linear combinations of others, the equations are called
linearly dependent, and they cannot be solved.
Figures 9.22 and 9.23 show the set of functions that convert the rst n rows and columns of
an array to an upper triangular form. These and other functions use a user dened type, status,
with possible values ERROR returned if there is an error, and OK returned otherwise. The type
status is dened as follows:
typedef enum {ERROR, OK} status
We also assume a maximum of MAX equations, so the two dimensional array must have MAX rows
and MAX+1 columns. Figure 9.21 includes the header le with the denes and function prototypes
used in the program. Since precision is important in these computations, we have used formal
parameters of type double. The two dimensional arrays can store coecients for a maximum of
MAX equations (rows) and have MAX + 1 columns to accommodate the right hand side values.
The function uptriangle() transforms coecients of the equations to an upper triangular
form. For each k from 0 through n-1, it calls findpivot() to nd the pivot in the kth column.
If no pivot is found, findpivot() will return an ERROR (findpivot() is called even for the
(n ; 1)st column even though there is no lower (n ; 1)st column to test if an-1]n-1] is zero). If
findpivot() returns OK, then uptriangle() calls process col() to reduce the lower k th column
to zero. We have included debug statements in process col() to help track the process. The
function pr2adbl() prints the two dimensional array | we will soon write this function.
The function findpivot() calls on function findnonzero() to nd a non-zero pivot in column
k if ak]k] is zero. If a pivot is found, it swaps the appropriate rows and returns OK. Otherwise,
it reurns ERROR. The function findnonzero() merely scans the lower column k for a non-zero
element. It either returns the row in which it nds a non-zero element or it returns -1 if no such
element is found. Rows of the array are swapped by the function swaprows() which also includes
a debug statement to prints the row indices of the rows being swapped.
When uptriangle() returns with OK status, the array will be in upper triangular form. The
next step in solving the equations is to employ back substitution to nd the values of the unknowns.
9.5. AN EXAMPLE: LINEAR ALGEBRAIC EQUATIONS 375
/* File: gauss.h */
typedef enum {ERROR, OK} status
#define DEBUG
#define MAX 10 /* maximum number of equations */
status uptriangle(double a]MAX + 1], int n)
void process_col(double a]MAX + 1], int k, int n)
status findpivot(double a]MAX + 1], int k, int n)
int findnonzero(double a]MAX + 1], int k, int n)
void swaprows(double a]MAX + 1], int k, int j, int n)
status gauss(double a]MAX + 1], double x], int n)
int getcoeffs(double a]MAX + 1])
void pr2adbl(double a]MAX + 1], int n)
void pr1adbl(double x], int n)
Figure 9.21: Header File for Gauss Functions
We now examine the back substitution process. As we saw earlier, we must solve equations
backwards starting at index n ; 1 and proceeding to index 0. The ith equation in upper triangular
form looks like this:
ai]i]*xi] + ai]i+1]*xi+1] + ... + ai]n-1]*xn-1] = ai]n]
Recall, in our representation, the right hand side is the nth column of the two dimensional array.
For each index i, we must sum all contributions from those unknowns already solved for, i.e. those
xi] with index greater than i. This is the following sum:
sum = ai]i+1]*xi+1] + ... + ai]n-1]*xn-1]
We then subtract this sum from the right hand side, ai]n], and divide the result by ai]i]
to determine the solution for xi]. The algorithm is shown below:
/* Algorithm: Back_Substitution */
for (i = n - 1 i >= 0 i--) { /* go backwards */
sum = 0
for (j = i + 1 j <= n - 1 j++) /* sum all contributions from */
sum += ai]j] * xj] /* xj] with j > i */
xi] = (ai]n] - sum) / ai]i] /* solve for xi] */
}
We can now write the function gauss() that solves a set of equations by the Gaussian elim-
ination method which rst calls on uptriangle() to convert the coecients to upper triangular
form. If this succeeds, then back substitution is carried out to nd the solutions. As with other
functions, gauss() returns OK if successful, and ERROR otherwise. The code is shown in Figure
9.24. The code is straight forward. It incorporates the back substitution algorithm after the
function call to uptriangle(). If the function call returns ERROR, the equations cannot be solved
376 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: gauss.c */
#include <stdio.h>
#include "gauss.h"
/* Implements the Gauss method to transform equations to
an upper triangular form.
*/
status uptriangle(double a]MAX + 1], int n)
{ int i, j, k
for (k = 0 k < n k++) {
if (findpivot(a, k, n) == OK)
process_col(a, k, n)
else
return ERROR
}
return OK
}
/* Zeros out coefficients in column k below the main diagonal. */
void process_col(double a]MAX + 1], int k, int n)
{ int i, j
double m
for (i = k + 1 i < n i++) {
m = -ai]k] / ak]k]
for (j = k j <= n j++)
ai]j] += m * ak]j]
#ifdef DEBUG
printf("Multiplier for row %d is %6.2f\n", i, m)
pr2adbl(a, n)
#endif
}
}
Figure 9.22: Code for Functions to do Gaussian Elimination
9.5. AN EXAMPLE: LINEAR ALGEBRAIC EQUATIONS 377
/* Finds a non-zero pivot element in column k and row with
index >= k.
*/
status findpivot(double a]MAX + 1], int k, int n)
{ int j
void swaprows()
if (ak]k] == 0) {
j = findnonzero(a, k, n)
if (j < 0)
return ERROR
else
swaprows(a, k, j, n)
#ifdef DEBUG
printf("Rows %d and %d swapped\n", k, j)
#endif
}
return OK
}
/* Scans the rows with index >= k for the first non-zero element
in the kth column of the array 'a' of size n.
*/
int findnonzero(double a]MAX + 1], int k, int n)
{ int i
for (i = k i < n i++)
if (ai]k])
return(i)
return(-1)
}
/* Swaps the kth and the jth rows in the array 'a' with n rows. */
void swaprows(double a]MAX + 1], int k, int j, int n)
{ int i
double temp
for (i = k i <= n i++) {
temp = ak]i]
ak]i] = aj]i]
aj]i] = temp
}
}
Figure 9.23: Code for Functions to do Gaussian Elimination | continued
378 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: gauss.c - continued */
/* Transforms equations to upper triangular form using Gauss
method. Then, solves equations, one at a time.
*/
status gauss(double a]MAX + 1], double x], int n)
{ int i, j
double sum
if (uptriangle(a, n) == ERROR) {
printf("Dependent equations - cannot be solved\n")
return ERROR
}
for (i = n - 1 i >= 0 i--) {
sum = 0
for (j = i + 1 j <= n - 1 j++)
sum += ai]j] * xj]
if (ai]i])
xi] = (ai]n] - sum) / ai]i]
else
return ERROR
}
return OK
}
Figure 9.24: Code for gauss()
9.5. AN EXAMPLE: LINEAR ALGEBRAIC EQUATIONS 379
and gauss() returns ERROR. Otherwise, gauss() proceeds with back substitution and stores the
result in the array x]. Since all ai]i] must be non-zero at this point, we do not really need
to test if ai]i] is zero before using it as a divisor
however, we do so as an added precaution.
We are almost ready to use the function gauss() in a program. Before we can do so
however,
we need some utility functions to read and print data. Here are the descriptions of these functions:
getcoeffs(): reads the coecients and the right hand side values into an array
it returns the
number of equations.
pr2adbl(): prints an array with n rows and n + 1 columns.
pr1adbl(): prints a solution array.
All these functions use data of type double. The code is shown in Figure 9.25.
Finally, we are ready to write a program driver as shown in Figure 9.26. The driver rst reads
coecients and the right hand side values for a set of equations and then calls on gauss() to
solve the equations. During the debug phase, both the original data and the transformed upper
triangular version are printed. Finally, if the equations are solved with success, the solution is
printed. Otherwise, an error message is printed. During debugging, the macro DEBUG is dened
in gauss.h so that we can track the process. The program loops as long as there are equations
to be solved. In each case, it gets coecients using getcoeffs() and solves them using gauss().
During debug, the program uses pr2adbl() to print the original array and the array after gauss
transformation. If the solution is possible, the program prints the solution array using pr1adbl().
Here are several example equation solutions:
Sample Session:
***Simultaneous Equations - Gauss Elimination Method***
Number of equations, zero to quit: 2
Type coefficients and right side of each row
Row 0: 132
Row 1: 352
Original equations are:
1.00 3.00 2.00
3.00 5.00 2.00
Multiplier for row 1 is -3.00
1.00 3.00 2.00
0.00 -4.00 -4.00
Equations after Gauss Transformation are:
1.00 3.00 2.00
0.00 -4.00 -4.00
Solution is:
-1.00
1.00
380 CHAPTER 9. TWO DIMENSIONAL ARRAYS
/* File: gauss.c - continued */
/* Function gets the coefficients and the right hand side of equations.
*/
int getcoeffs(double a]MAX + 1])
{ int i, j, n
printf("Number of equations, zero to quit: ")
scanf("%d", &n)
if (n)
printf("Type coefficients and right side of each row\n")
for (i = 0 i < n i++) {
printf("Row %d: ", i)
for (j = 0 j <= n j++)
scanf("%lf", &ai]j])
}
return n
}
/* Prints coefficients and right side of equations */
void pr2adbl(double a]MAX + 1], int n)
{ int i, j
for (i = 0 i < n i++) {
for (j = 0 j <= n j++)
printf("%10.2f ", ai]j])
printf("\n")
}
}
/* Prints the solution array */
void pr1adbl(double x], int n)
{ int i
for (i = 0 i < n i++)
printf("%10.2f\n", xi])
}
Figure 9.25: Code for Utility Functions for Gauss Program
9.5. AN EXAMPLE: LINEAR ALGEBRAIC EQUATIONS 381
/* File: gauss.c
Header Files: gauss.h
This program solves a number of simultaneous linear algebraic
equations using the Gauss elimination method. The process repeats
itself until number of equations is zero.
*/
main()
{ double aMAX]MAX + 1] /* coefficients and right hand side */
double xMAX] /* solution */
int n /* number of equations */
status soln /* status of solution, OK or ERROR */
printf("***Simultaneous Equations***\n\n")
while (n = getcoeffs(a)) {
printf("\nOriginal equations are:\n")
#ifdef DEBUG
pr2adbl(a, n)
#endif
soln = gauss(a, x, n)
#ifdef DEBUG
printf("\nEquations after Gauss Transformation are:\n")
pr2adbl(a, n)
#endif
if (soln == OK) {
printf("\nSolution is:\n")
pr1adbl(x, n)
}
else printf("Equations cannot be solved\n")
}
}
Figure 9.26: Driver Program for Gaussian Elimination
382 CHAPTER 9. TWO DIMENSIONAL ARRAYS
Number of equations, zero to quit: 3
Type coefficients and right side of each row
Row 0: 1234
Row 1: 4321
Row 2: 0725
Original equations are:
1.00 2.00 3.00 4.00
4.00 3.00 2.00 1.00
0.00 7.00 2.00 5.00
Multiplier for row 1 is -4.00
1.00 2.00 3.00 4.00
0.00 -5.00 -10.00 -15.00
0.00 7.00 2.00 5.00
Multiplier for row 2 is -0.00
1.00 2.00 3.00 4.00
0.00 -5.00 -10.00 -15.00
-0.00 7.00 2.00 5.00
Multiplier for row 2 is 1.40
1.00 2.00 3.00 4.00
0.00 -5.00 -10.00 -15.00
0.00 0.00 -12.00 -16.00
Equations after Gauss Transformation are:
1.00 2.00 3.00 4.00
0.00 -5.00 -10.00 -15.00
0.00 0.00 -12.00 -16.00
Solution is:
-0.67
0.33
1.33
Number of equations, zero to quit: 3
Type coefficients and right side of each row
Row 0: 1234
Row 1: 2468
Row 2: 3179
Original equations are:
1.00 2.00 3.00 4.00
2.00 4.00 6.00 8.00
3.00 1.00 7.00 9.00
Multiplier for row 1 is -2.00
9.6. COMMON ERRORS 383
1.00 2.00 3.00 4.00
0.00 0.00 0.00 0.00
3.00 1.00 7.00 9.00
Multiplier for row 2 is -3.00
1.00 2.00 3.00 4.00
0.00 0.00 0.00 0.00
0.00 -5.00 -2.00 -3.00
Rows 1 and 2 swapped
Multiplier for row 2 is 0.00
1.00 2.00 3.00 4.00
0.00 -5.00 -2.00 -3.00
0.00 -0.00 -0.00 -0.00
Dependent equations - cannot be solved
Equations after Gauss Transformation are:
1.00 2.00 3.00 4.00
0.00 -5.00 -2.00 -3.00
0.00 -0.00 -0.00 -0.00
Equations cannot be solved
Number of equations, zero to quit: 0
The rst two sets of equations are solvable
the last set is not because the second equation
in the last set is a multiple of the rst equation. Thus these equations are linearly dependent
and they cannot be solved uniquely. In this case, after the zeroth lower column is reduced to
zero, a1]1] is zero. A pivot is found in row 2, rows 1 and 2 are swapped, and lower column
1 is reduced to zero. However, a2]2] is now zero, and there is no unique way to solve these
equations.
If the coecients are such that the equations are almost but not quite linearly dependent,
the solution can be quite imprecise. An improvement in precision may be obtained by using an
element with the largest absolute value as the pivot. Implementation of an improved version of
the method is left as an exercise.
9.6 Common Errors
1. Failure to specify ranges of smaller dimensional arrays in declaration of formal parameters.
All but the range of the rst dimension must be given in a formal parameter declaration.
Example:
init2(int aray2]])
{
...
}
Error! aray2 is a pointer to a two dimensional array, i.e. it points to an object that is a
one-dimensional array, aray20]. Without a knowledge of the size of the object, aray20],
384 CHAPTER 9. TWO DIMENSIONAL ARRAYS
it is not possible to access aray21], aray22], etc. Consequently, one must specify the
number of integer objects in aray20]:
init2(int aray2]COLS])
{ ...
}
Correct! aray20] has COLS objects. It is possible to advance the pointer, aray2 correctly
to the next row, etc.
2. Failure to pass arguments correctly in function calls:
init2(aray2MAX]COLS])
init2(aray2]COLS])
init2(aray2]])
All of the above are errors. A two dimensional array name is passed in a function call:
init2(aray2)
3. Confusion between pointers to dierent types of objects. For example, in the above, aray2
points to an array object, aray20], whereas aray20] points to an int object. The
expression aray2 + 1 points to aray21], whereas aray20] + 1 points to aray20]1].
In the rst case the pointer is increased by COLS integer objects, whereas in the second case
the pointer is increased by one integer object.
4. Confusion between arrays of character strings and arrays of character pointers:
char tableMAX]SIZE], *ptrarayMAX]
The rst declares table to be a two dimensional array that can be used to store an array of
strings, one each in table0], table1], tablei], etc. The second declares ptraray to be
an array, each element of which is a char *. Read the declaration from the end: MAX] says
it is an array with MAX elements
ptraray is the name of the array
char * says each element
of the array is a char *. Properly initialized with strings stored in table]], tablei] can
point to a string. Properly initialized with pointers to strings, ptrarayi] can also point
to a string. However, tableMAX]SIZE] provides memory space for the strings, whereas
ptrarayMAX] provides memory space only for pointers to strings. Both pointers may be
used in a like manner:
puts(tablei])
puts(ptrarayi])
They will both print the strings pointed to by the pointers.
9.7. SUMMARY 385
9.7 Summary
In this chapter we have seen that, in C, the concept of an array can be extended to arrays of
multi-dimensions. In particular, a two dimensional array is represented as a one dimensional
array, each of whose elements, themselves, are one dimensional arrays, i.e. and array or arrays.
Similarly, a three dimensional array is an array whose elements are each 2 dimensional arrays (an
array of arrays of arrays). We have seen how such arrays are declared within programs and how
they are organized in memory (row major order). We have seen how we can access the elements
of multi dimensional arrays using the subscripting notation and the correspondence between this
notation and pointer values. Because for higher dimensional arrays, the pointer expressions may
get complicated and confusing, in general, most programs use the subscripting notations for arrays
of two dimensions or more. We have also shown that when passing arrays to functions, the size of
all dimensions beyond the rst must be specied in the formal parameter list of the function so
that the location of all elements can be calculated by the compiler.
Throughout the chapter we have seen applications for which a two dimensional data structure
provides a convenient and compact way of organizing information. These have included data base
applications, such as our payroll and student test score examples, as well as using two dimensional
arrays to store an array of strings. We have seen how we can then use this later data structure to
search and sort arrays of strings, and have shown that for this data type, as well as other large
data types, it is often more ecient to work with arrays of pointers when reordering such data
structures.
Finally, we have developed a rather large application using 2D arrays | solutions to simulta-
neous linear equations usinf Gaussian elimination. This is one algorithm for doing computations
in the realm of linear algebra
several additional examples common in engineering problems are
presented in Chapter 15.
One last point to remember about multi-dimensional arrays: this data structure is a very useful
way to organize a large collection of data into one common data structure
however, all of the
data items in this structure must be of the same type. In the next chapter we will see another
compound data type provided in C which does not have such a restriction | the structure.
386 CHAPTER 9. TWO DIMENSIONAL ARRAYS
9.8 Exercises
Given the following declaration:
int x10]20]
Explain what each of the following represent:
1. x
2. x + i
3. *(x + i)
4. *(x + i) + j
5. *(*(x + i) + j)
6. x0]
7. xi]
8. xi] + j
9. *(xi] + j)
Find and correct errors if any. What does the program do in each case?
10. main()
{ int x5]10]
init(x]])
}
void init(int a]])
{ int i, j
for (i = 0 i < 10 i++)
for (j = 0 j < 5 j++)
ai]j] = 0
}
11. main()
{ int x5]10]
init(x]])
}
void init(int *a)
{ int i, j
9.8. EXERCISES 387
for (i = 0 i < 10 i++)
for (j = 0 j < 5 j++) {
*a = 0
a++
}
12. main()
{ char s5]100]
read_strings(s)
print_strings(s)
}
read_strings(char s]100])
{
for (i = 0 i < 5 i++) {
gets(*s)
s++
}
}
print_strings(char s]100])
{
while (*s) {
puts(s)
s++
}
}
388 CHAPTER 9. TWO DIMENSIONAL ARRAYS
9.9 Problems
1. Read id numbers, project scores, and exam scores in a two dimensional array from a le.
Compute the averages of each project and exam scores
compute and store the weighted
average of the scores for each id number.
2. Repeat 1, but sort and print the two dimensional array by weighted average in decreasing
order. Sort and print the array by id numbers in increasing order. Use an array of pointers
to sort.
3. Repeat 2, but plot the frequency of each weighted score.
4. Combine 1-3 into a menu-driven program with the following options: read names, id numbers,
and scores from a le
add scores for a new project or exam
save scores in a le
change
existing scores for a project or an exam for specied id numbers
delete a data record
add
a data record
compute averages
sort scores in ascending or descending order by a primary
key, e.g. id numbers, weighted scores, etc.
compute weighted average
plot frequency of
weighted scores
help
quit.
5. Write a function that uses binary search algorithm to search an array of strings.
6. Write a function that sorts strings by selection sort in either increasing or decreasing order.
7. Write a program that takes a string and breaks it up into individual words and stores them.
8. Repeat 7 and keep track of word lengths. Display the frequency of dierent word lengths.
9. Repeat 7, but store only new words that occur in a string. If the word has already been
stored, ignore it.
10. Write a function that checks if the set of words in a string, s, represents a subset of the set
of words in a second string, t. That is, the words of s are all contained in t, with t possibly
containing additional words.
11. Write a menu-driven spell check program with the following options: read a dictionary from
a le
spell check a text le
add to dictionary
delete from dictionary
display text buer
save text buer
help
quit.
The dictionary should be kept sorted at all times and searched using binary search. Use
an array of pointers to sort when new entries are inserted. In the spell check option, the
program reads in lines of text from a le. Each word in a line is checked with the dictionary.
If the word is present in the dictionary, it is ignored. Otherwise, the user is asked to make
a decision: replace the word or add it to the dictionary. Either replace the word with a new
word in the line or add the word to dictionary. Each corrected line is appended to a text
buer. At the quit command, the user is alerted if the text buer has not been saved.
12. Write a simple macro processor. It reads lines from a source le. Ignoring leading white
space, each line is examined to see if it is a control line starting with a symbol # and followed
by a word "dene". If it is, store the dened identier and the replacement string. Each
line is examined for the possible occurrence of each and every dened identier
if a dened
9.9. PROBLEMS 389
identier occurs in a line, replace it with the replacement string. The modied line must be
examined again to see if a dened identier exists
if so, replace the identier with a string,
etc.
13. Write a lexical scanner, scan(), which calls nexttok() of Problem 42 to get the next token
from a string. Each new token or symbol of type identier, integer, or oat that is found
is stored in a symbol table. The token is inserted in the table if and only if it was not
already present. A second array keeps track of the type of each token stored at an index.
The function scan() uses srcharay() to search the array of tokens, uses inserttok() to
insert a token in an array, and uses inserttype() to insert type of a token.
The function scan() returns a token in a string, type of the token, and index where the
token is stored in the array. If the array is lled, a message saying so must be displayed.
Write a program driver to read strings repeatedly. For each string, call scan() to get a
token. As scan() returns, print the token, its type, and index. Repeat until an end of string
token is reached. When the end of le is encountered, print each of the tokens, its type, and
index.
14. Write routines for drawing lines and rectangles. Write a program that draws a specied
composite gure using a character '*'. Allow the user to specify additions to the gure and
display the gure when the user requests.
15. Modify 14 to a menu-driven program that allows: draw horizontal and vertical lines, hori-
zontally oriented rectangles, lled rectangles, display gure, help, and quit.
16. Write a program that plays a game of tic-tac-toe with the user. The game has three rows
and three columns. A player wins when he succeeds in lling a row or a column or a diagonal
with his mark, '0'. The program uses '*'. Write and use the following functions:
init board(): initialize the board
display board(): dispays the board
enter move(): for user to enter a move in row and col
state of game(): test state, nish or continue
17. Modify the Gauss Method so that a pivot with the largest magnitude is used in converting
the array of coecients to an upper triangular form.
18. Modify 17 to a menu-driven program that allows the following commands: Get coecients,
display coecients, solve equations, display solution, verify solution, help, and quit. Write
and use functions get coeffs(), display coeffs(), solve eqns(), display soln(), verify soln(),
help().
19. Modify 18 so that the input data is in the form:
a00 x0 + a01 x1 + a02 x2 = b1
20. Modify 19 so that display coecients displays equations in the above form.
390 CHAPTER 9. TWO DIMENSIONAL ARRAYS
21. Write a simple menu driven editor which allows the following commands: text insert, display
text, delete text, delete lines, insert lines, nd string, nd word, replace string, replace word,
help, and quit. A window should display part of the text when requested.
PART II
391
392 CHAPTER 9. TWO DIMENSIONAL ARRAYS
Chapter 10
Sorting and Searching
One very common application for computers is storing and retrieving information. For example,
the telephone company stores information such as the names, addresses and phone numbers of its
customers. When you dial directory assistance to get the phone number for someone, the operator
must look up that particular piece of information from among all of data that has been stored.
Taken together, all of this information is one form of a data base which is organized as a collection
of records. Each record consists of several elds, each containing one piece of information, such as
the name, address, phone number, id number, social security number or part number, etc..
As the amount of information to be stored and accessed becomes very large, the computer
proves to be a useful tool to assist in this task. Over the years, as computers have been applied to
these types of tasks, many techniques and algorithms have been developed to eciently maintain
and process information in data bases. In this chapter, we will develop and implement some of
the simpler instances of these algorithms. The processes of \looking up" a particular data record
in the data base is called searching. We will look at two dierent search algorithms
one very
easy to implement, but inecient, the other much more ecient. As we will see, in order to do
an ecient search in a data base, the records must be maintained in some order. For example,
consider the task of nding the phone number in the phone book of someone whose name you
know, as opposed to trying to nd the name of someone whose phone number you know in the
same book.
The process of ordering the records in a data base is called sorting. We will discuss three sorting
algorithms and their implementation in this chapter, as well. Sorting and searching together
constitute a major area of study in computational methods. We present some of these methods
here to introduce this area of computing as well as to make use of some of the programming
techniques we have developed in previous chapters.
As we develop these algorithms, we use a very simple data base of records consisting of single
integers only. We conclude the chapter by applying the searching and sorting techniques to our
payroll data base with records consisting of multiple numeric elds. In Chapter 9 we will see how
these same algorithms can be applied to string data types described in Chapter 11.
10.1 Finding a Data Item | The Search Problem
Suppose we have a collection of data items of some specic type (e.g. integers), and we wish to
determine if a particular data item is in the collection. The particular data item we want to nd
393
394 CHAPTER 10. SORTING AND SEARCHING
key == ?
? ? ? ?
index 0 1 2 MAX
Figure 10.1: Sequential Search
is called the key and out task is to search the records in the data base to nd one which \matches"
the key.
The rst decision we must make is how to represent the collection of data items. In Chapter
7 we saw a data structure which could hold a collection of data items all of the same type: the
array. So we can consider our data base of integer values to be stored in an array. Our task then
becomes:
Task:
SRCH0: Search an array for an index where the key is located
if key is not present, print a
message. Repeat until an end of le is entered for the key.
In this task, we choose to return the index where the key is located because this index will
allow us to retrieve the entire record in the case where our array is part of a database. The
simplest approach to determine if a key is present in an array is to make an exhaustive search of
the array. Start with the rst element, if the key matches, we are done
otherwise move on to the
next element and compare the key, and so on. We simply traverse the array in sequence from the
rst element to the last as shown in Figure 10.1. Each element is compared to the key. If the
key is found in the array, the corresponding array index is returned. If the item is not found in
the array, an invalid index, say -1, is returned. This type of search is called Sequential Search or
Linear Search because we sequentially examine the elements of the array. In the worst case, the
number of elements that must be compared with the key is linearly proportional to the size of the
array.
Linear search is not the most ecient way to search for an item in a collection of items
however,
it is very simple to implement. Moreover, if the array elements are arranged in random order, it is
the only reasonable way to search. In addition, eciency becomes important only in large arrays
if the array is small, there aren't many elements to search and the amount of time it takes is not
even noticed by the user. Thus, for many situations, linear search is a perfectly valid approach.
Here is a linear search algorithm which returns the index in the array where key is found or -1 if
key is not found in the array:
initialize index i to 0
traverse the array until exhausted
if arrayi] matches key
return i
return -1.
You might also like
- Growth Chamber in Biochemistry TC2ManualNo ratings yetGrowth Chamber in Biochemistry TC2Manual76 pages
- Advanced Visual Basic 6 Power Techniques For Everyday Programs.9780201707120.24922100% (2)Advanced Visual Basic 6 Power Techniques For Everyday Programs.9780201707120.24922515 pages
- At Full Speed With Python: João VenturaNo ratings yetAt Full Speed With Python: João Ventura21 pages
- Data Structure - SEM III - Aeraxia - inNo ratings yetData Structure - SEM III - Aeraxia - in98 pages
- Data Structure and Algorithms (2 Files Merged)No ratings yetData Structure and Algorithms (2 Files Merged)107 pages
- M.Rajanikanth Lecturer in Computer Science DRG GDC PentapaduNo ratings yetM.Rajanikanth Lecturer in Computer Science DRG GDC Pentapadu13 pages
- Arrays in C - Programmerdouts: Why There Is A Need of Array ?No ratings yetArrays in C - Programmerdouts: Why There Is A Need of Array ?3 pages
- Senior Five Mce MPC Notes Data Structure and AlgorithmsNo ratings yetSenior Five Mce MPC Notes Data Structure and Algorithms122 pages
- C Arrays: - A Collection of Same Type Data, 1D, 2DNo ratings yetC Arrays: - A Collection of Same Type Data, 1D, 2D25 pages
- Chapter Six: Introduction To Arrays, String and PointersNo ratings yetChapter Six: Introduction To Arrays, String and Pointers44 pages
- Student Solutions Manual for Mathematics for Economics, fourth editionFrom EverandStudent Solutions Manual for Mathematics for Economics, fourth editionNo ratings yet
- OSP-P300/P300A Series: Operation ManualNo ratings yetOSP-P300/P300A Series: Operation Manual32 pages
- JAVA Sample Questions For Practice (II CSE - A' & II IT - B')No ratings yetJAVA Sample Questions For Practice (II CSE - A' & II IT - B')5 pages
- (Ebook) Mastering C++ Programming Language: A Beginner’s Guide by Sufyan Bin Uzayr, (ed.) ISBN 9781003214762, 9781032103204, 1032103205, 1003214762 - The complete ebook version is now available for download100% (3)(Ebook) Mastering C++ Programming Language: A Beginner’s Guide by Sufyan Bin Uzayr, (ed.) ISBN 9781003214762, 9781032103204, 1032103205, 1003214762 - The complete ebook version is now available for download83 pages
- Unit - 1: 1) What Is Object Oriented Programming? What Are The Benefits and Applications of Oop?No ratings yetUnit - 1: 1) What Is Object Oriented Programming? What Are The Benefits and Applications of Oop?51 pages
- Fbs-30Gm Fbs-30Gm Motion Controller User Manual: Fatek Automation CorporationNo ratings yetFbs-30Gm Fbs-30Gm Motion Controller User Manual: Fatek Automation Corporation185 pages
- Attendence Tracker CS For Class 12TH Project.No ratings yetAttendence Tracker CS For Class 12TH Project.15 pages
- Release Notes - 3BDS007582D101 - Application BuilderNo ratings yetRelease Notes - 3BDS007582D101 - Application Builder33 pages
