ORACLE 11G(grid)

Data: 27/2/2015

There are mainly two different types of database language is ORACLE. They are:

1) SQL (STRUCTURED QUERY LANGUAGE)

2) PL/SQL (PROCEDURAL LANGUAGE extension of SQL)
3) Dynamic SQL.(optional)

ORACLE is a Relational Database product which is used to store data permanently in

secondary storage devices. If you want to operate ORACLE then we are using following
languages:

1) SQL: It is non-procedural language.

2) PL/SQL: It is a procedural language.

All organizations store same type of data.

DATA: It is a collection of Raw facts.

Example: 101 dinesh 2000

102 mahesh 3000

In above example, there no meaningful data such data is known as Raw facts.

INFORMATION: It is a collection of meaningful data or processed data.

Example: EmpID Ename Salary

101 dinesh 2000

102 mahesh 3000

In the above example, there is meaningful data which is in table format which consist of
three different fields.

DATA STORE: It is a place where we can store data or information.

1) Books & Papers
2) Flat files
3) Database
FLAT FILES: This is a traditional mechanism which is used to store data or information in
individual unrelated files. These files are also called as Flat Files.
Drawbacks of Flat files:
1) Data Retrieval
2) Data Redundancy
3) Data Integrity

MUARLI SIR 2015

1
4) Data Security
5) Data Indexing

1) Data Retrieval: If we want to retrieve data from flat files then we must develop
application program in high level languages, where as if we want to retrieve data from
databases then we are using Sequel Language.

SEQUEL (Structured English Query Language)

Date: 28/2/2015

2) Data Redundancy: Sometimes we are maintaining multiple copies of the same data in
different locations this data is also called as Duplicate data or Redundant data. In Flat files
mechanism when we are modifying data in one location it is not effected in another
location. This is called INCONSISTENCY.

Date: 2/3/2015

In databases, every transaction internally having 4 properties. These properties are known as
ACID properties.
ACID Properties:
A mean Atomicity (ROLLBACK)
C mean Consistency
I mean Isolation
D mean Durability (COMMIT)
These properties only automatically maintains consistent data in databases.
3) Data Integrity: Integrity means to maintain proper data. If we want to maintain proper
data then we are defining set of rules, these rules are also called as “ Business rules”. In
databases, we are maintaining proper data using „constraints‟, „triggers‟. If we want to
maintain proper data in flat files we must develop application programs in high level
languages like COBOL, JAVA, ETC…..
4) Data Security: Data stored in flat files cannot be secured because flat files doesn‟t
provides security mechanism. Whereas databases provides “ROLE based security”.
5) Data Indexing: If we want to retrieve data very quickly from database then we are using
indexing mechanism. Whereas flat files doesn‟t provide indexing mechanism.

To overcome all the above problems, a new software used by all organization to store data or
information in secondary storage devices. This is called DBMS software.

DBMS (DATABASE MANAGEMENT SYSTEM):

It is a collection of programs (S/W) written to manage database.

Example: ORACLE, FOXPRO, DB2, TERADATA, SQLSERVER, SYBASE, MYSQL, INGRESS,

INFORMIX, SQLLITE…… etc;

MUARLI SIR 2015

2
In file based approach, every application program in the organization maintain its own file
separate from other application program.

FILE based approach:

Here custno and cust name attributes are duplicate data.

THIS IS OLD APPROACH

Once we are installing DBMS software into our system then automatically some place is
created in hard disc. This is called “Physical Database”. And also automatically an user
interface is created. Through the user interface we can also directly interactive with the
database or indirectly interactive with the database using application programs in high level
languages.

MUARLI SIR 2015

3
“AFTER INSTALLING DBMS SOFTWARE”

Date: 3/3/2015

DATABASE: It is an organized collection of interrelated data used by application program in

an organization. Once data stored in database it can be shared by number of users
simultaneously and also this data can be integrated.

DBMS Architecture: American National Standard Institute(ANSI) has established three level
architecture for database. This architecture is also called as “ansi/sparc” (Standard Planning
And Requirements Committee) architecture.

Main objective of DBMS architecture is to separate users view of the database from the where
physically it is stored.

This architecture mainly consist of three levels. They are:

1) External level
2) Conceptual level
3) Internal level

MUARLI SIR 2015

4
DBMS Architecture or Three Level Architecture

DBMS architecture provides “DATA INDEPENDENCE”.

Data Independence: Upper levels are unaffected by changes in the lower levels is called as
“Data Independence”. DBMS architecture have two types of Data Independences:

1) Logical Data Independence

2) Physical Data Independence
1) Logical Data Independence: Changes to the conceptual level do not required to change
to the external level this is called Logical Data Independence.
Example: Adding a new entity in conceptual level does not effect in external level.
2) Physical Data Independence: Changes to the internal level do not required to changes
in conceptual level. This is called Physical Data Independence.
Example: Adding an index to the internal level it is not affected in conceptual level.

CONCEPTUAL LEVEL: It describes logical representation of the database. Conceptual level

defines type of data storing in database and also defines what type of data does not store in
database using constraints and also specifies the relationship between data items.

NOTE: This level does not define how data is stored in database. In relational databases we
are defining conceptual levels through tables.

EXTERNAL LEVEL: This level describes end user view of the database. i.e., in this level
some group of users access only part of the database. In this level only we are defining the
view and those views given to the number of users.

MUARLI SIR 2015

5
INTERNAL LEVEL: Internal level describes how physically data is stored in database. This
level is handled by database administrator only. In relational databases indexes, cluster are
available in internal level.

Data: 4/3/15

DATA MODELS: How data is represented at the conceptual level defined by means of “Data
Model” in the history of database design three data models have been used.

1) Hierarchical Data Model

2) Network Data Model
3) Relational Data Model

DBMS ARCHITECTURE

Hierarchical Data Model: This model introduced in 1960. In this data model organizes data in
tree like structure, we are representing data in parent child hierarchy. In this data model
also data is represented in the format of records and also record type is also same as table
in relational data model.
MUARLI SIR 2015
6
 This Data model having more duplicate records because this data model is implemented
based on one- to – many relationships. That is why in this data model always child segments
are repeated.

In this data model products, we are retrieved data very slowly because in this data
model products data base servers searching data based on root node onwards.

In 1960, IBM introduced IMS( Information Management System) product based on

Hierarchical Data Model.

If we want to operate hierarchical data model products then we are using procedural
language.

HIERARCHICAL DATA MODEL

RELATIONAL DATA MODEL

MUARLI SIR 2015

7
Date: 5/3/15

Network Data Model: In 1970‟s, CODASYL(Conference Or Data System Language)

committee introduced network data model. This data model is implemented based on many
– to – many relationships. In this data model also data is stored in format of records. And
also records type is also same as table in Relational Data model. In 1970‟s IBM company
introduced IDMS(Information Data Management System) based on network data model. If
we want to operate network data model products then we must use procedural language.

NETWORK DATA MODEL

Relational Data Model: In 1970, E.F.Codd introduced Relational Data Model. In this data
model we are storing data in 2- dimensional tables. Relational data model mainly consist of
3 components.

1) Collection of Objects.
2) Set of Operators.
3) Set of Integrity rules.

If we want Relational Data Model products then we are using Non- Procedural Language
“SQL”.

MUARLI SIR 2015

8
 In 1977 Larry Ellision, Bob Miner, Ed Oates founded new company “ Software
Development Laboratories” (SDL)
 In 1978 SDL Introduced oracle version1 (never released).
 In 1979 SDL name changed into RSI(Relational Software Incorporation).
 In 1982 RSI name changed into Oracle Corporation.

ORACLE VERSIONS
1)Oracle 2.0 -> 1979
-> First public release
-> In this 2.0 only basic SQL functionality is there “joins”.
2) Oracle 3.0 -> 1983
-> Rewritten in C language.
-> Commit, Roll back.
3) Oracle 4.0 -> 1984
-> Read Consistancy
-> exp/imp utility programs [export/import].
4) Oracle 5.0 -> 1985
-> Client server architecture.
5) Oracle 6.0 -> 1988
-> Introduced PL/SQL
-> Row level locks.
6) Oracle 7.0 -> 1992
-> Roles
-> Integrity constraints.
-> Stored Procedures
-> Stored Functions
-> Packages
-> Triggers
-> Data type “varchar” changed into “varchar2”
-> Truncate table.
7) Oracle 7.1 -> 1994
-> Introduced dynamic SQL
-> ANSI/ISO SQL-92.
8) Oracle 7.2 -> 1995
-> Inline views or sub queries used in from clause.
-> ref cursor (cursor variable)
9) Oracle 7.3 -> 1996
-> Bit map indexes.
-> utl_file package.
10) Oracle 8.0 -> 1997
-> Object technology
-> Columns increased per a table upto “1000”.
-> nested table, varray.
-> Instead of triggers.
MUARLI SIR 2015
9
11) Oracle 8i (i- internet) -> 1999
-> Materialized views.
-> Function based indexes
-> Case conditional statements.
-> Analytical functions.
-> Autonomous trasactions
-> rollup, cube.
-> BULK BIND
12) Oracle 9i -> 2001
-> 9i joins or ansi joins
-> merge statements.
-> multi table insert
-> flash back queries.
-> Renaming a column
13) Oracle 10g (g- grid) -> 2003
-> recycle bin
-> flash back table
-> indices of clause
-> regular expressions
-> wm_concat().
14) Oracle 11g -> 2007
-> Introduced continue statement in PL/SQL loops
-> Read only tables
-> Virtual Columns
-> Pivot() fuction.
-> Compound trigger
-> enable, disable clauses used in trigger specification
-> follow clause
-> Sequences used in PL/SQL without using dual table.
-> Named, mixed notations are used in a subprogram executed used select statement.
Date: 6/3/15

MUARLI SIR 2015

10
SQL (STRUCTURE QUERY LANGUAGE)
SQL is an Non- procedural language, which is used to operate all relational database
products.
 In 1970, E.F.Codd introduced Relational Data Model. And also DSL/ Alpha language. In
IBM , “System/R” team modified DSL/Alpha language into “SQUARE”. Again IBM
company change SQUARE into SEQUEL (Structured English Query Language). Then only
SEQUEL became “SQL”.
 In 1986 introducing ANSI in SQL
 In 1987 introducing ISO in SQL
 In 1989 releasing first version of ANSI/ISO SQL 89 that is SQL 89.
 In 1992 ANSI/ISO SQL92 -> SQL92.
 In 1999 ANSI/ISO SQL 99 -> SQL99.
 In 2003 ANSI/ISO SQL 03 -> SQL03.

SQL Sub Languages in every database:

1) Data Definition Language (DDL):

 Create
 Alter
 Drop
 Truncate
 Rename(oracle 9i)
2) Data Manipulation Language (DML):
 Insert
 update
 delete
 merge(oracle 9i)
3) Data Retrieval Language(DRL) or Data Query Language(DQL):
 SELECT
4) Transactional Control Language (TCL):
 Commit
 Rollback
 Savepoint
5) Data Control Language(DCL):
 Grant
 Revoke

Oracle 10g, 11g, 12c.(Enterprise Edition)

Username: scott
Password: tiger
Error: Account locked message displayed for first time. So, again we have to connect.
Username: \ sys as sysdba
Password: sys
SQL> alter user scott account unlock:
SQL> conn scott/tiger
Enter password: tiger
MUARLI SIR 2015
11
Confirm password: tiger

Date: 7/3/2015

Data Types: Data types specifies type of data within a table column Oracle have following
data types:
1. Number(p,s)
2. Char varchar2(maxsize)
3. Date.
1) Number(p,s):
p precision (total number of digits)
s scale ( it is used to store fixed, floating point numbers).
Syntax: columnname number(p,s)
Example: SQL> create table test(sno number(7,2));
SQL> insert into test values (12345.67);
SQL> select * from test;
Sno
---------
12345.67
SQL> insert into test values(123456.7)
Error: value larger than specified precision allows for this column.
Note: whenever we are using number(p,s) then total number of digits before decimal places
upto “p-s” number of digits.
[eg:- p-s => 7-2=5]
After Decimal:
SQL> insert into test values(12345.6789);
SQL> select * from test;
Sno
-------------
12345.68
SQL> insert into test values(12345.6725);
SQL> select * from test;
Sno
------------
12345.67
Note: whenever we are using number(p,s) and also if we are passing more number of digits
after decimal place then oracle server only automatically rounded that number based on
specified “scale”.
Number(p): It is used to store fixed numbers only. Maximum length of the precision is up to
“38”.
Example: SQL> create table test1(sno number(7));
SQL> insert into test1 values(99.9);
SQL> select * from test1;
Sno
MUARLI SIR 2015
12
-------
100
SQL> insert into test1 values(99.4);
SQL> select * from test1;
Sno
-----------
99
2) Char: It is used to store fixed length “alpha numeric” data in bytes. Maximum limit is
upto 2000 bytes.
Syntax: columnname char (size);
By default character data type having one byte. Character data type automatically
“Blank Padded”.
Blank Padded: Whenever we are passing less number of characters than the specified
data type size then oracle server automatically allocates blank spaces after the value.
Example:

Varchar2(maxsize): It is used to store variable length alphanumeric data in bytes.

Maximum limit is upto 4,000 bytes. Whenever we are using varchar2() data type oracle
server not blank padded. After the value passed into that column.
Oracle 7.0 introduced “varchar2()” data type prior to oracle 7.0 we are using
varchar2 data type. It is also same as varchar2() data type. But it will stores upto 2000
bytes.
Syntax: columnname varchar2(maxsize);

3) Date: It is used to store dates in oracle date format.

Syntax: columnname date;
In oracle by default date format is
DD- MON-YY

Date: 9/3/15
Data Definition Language(DDL):
 Create
 Alter
 Drop

MUARLI SIR 2015

13
 Truncate
 Rename (orace 9i)
These commands are used to define structure of the table.
1) Create: It is used to create database objects like tables, views, sequences, indexes.
Creating a table:
Syntax: create table tablename(columnname1 datatype(size), columnname2
datatype(size)….. );
Example: SQL> create table first(sno number(10), name varchar2(10));
To view structure of the table:
Syntax: desc tablename;
Example: SQL> desc first;
2) Alter: It is used to change structure of the existing table.

Add: It is used to add number of columns into the existing table.

Syntax: Alter table tablename add(col1 datatype(size));
Example: SQL> alter table first add sal number(10);
SQL> desc table;

Modify: It is used to change column datatype or column datatype size only.

Syntax: alter table tablename modify(col datatype(size), col2 datatype(size));
Example: SQL> alter table first modify sno date;

Drop: It is used to remove columns from the existing table.

Method1: If we want to drop a single column at a time without using parenthesis”()” then
we are using following syntax:
Alter table tablename drop column columnname;
Example: SQL> alter table first drop column sal;
SQL> desc first;
Method2: If we want to drop single or multiple columns with using parenthesis then we are
using following syntax.
Syntax: alter table tablename drop(colname1, colname2, colname3….);
Example: SQL> alter table first drop(sal);
Important:
SQL> alter table first drop column sno;
Error: cannot drop all columns in a table.
NOTE: In all database systems we cannot drop all columns in a table.
3) DROP: It is used to remove database objects from database.
Syntax: drop objecttype objectname;
1. Drop table tablename;
2. Drop view viewname;

MUARLI SIR 2015

14
Droping a Table:

Before Oracle 10g:

Syntax: drop table tablename;

Oracle 10g (Enterprise Edition)

Syntax: drop table tablename;

Get it back from recycle bin.

Syntax: flashback table tablename to before drop;
To drop permanently:
Syntax: drop table tablename purge;
Example: (dropping a table  oracle 10g Enterprise Edition)
SQL> drop table first;
Table dropped.
Get it back the table:
SQL> flashback table first to before drop;

Testing:
Method1: SQL> desc first;
Error: Object first does not exist.

Method2: SQL> flashback table first to before drop;

Error: object not in Recycle bin.

Date: 10/3/15

RECYCLE BIN: Oracle 10g introduced recycle bin which is used to store dropped tables.
Recycle bin is a read only table whenever we are installing oracle then automatically so
MUARLI SIR 2015
15
many read only tables are created. These read only tables are also called as “ Data
Dictionaries”.

In oracle, we can also drop tables from recycle bin using “purge” command.
To Drop particular table from Recycle bin:
Syntax: purge table tablename;
Example: SQL> create table first(sno number(10));
SQL> drop table first;
SQL> desc recycle bin;
SQL> select original_name from recyclebin;
Original_name
--------------------
First
SQL> purge table first;
Testing:
SQL> select original_name from recyclebin;
No rows selected;

To Drop all tables from Recycle bin:

Syntax: SQL> purge recyclebin;
Example: SQL> create table first(sno number(10));
SQL> create table second(sno number(10));
SQL> drop table first;
SQL> drop table second;
SQL> desc recyclebin;
SQL> select original_name from recyclebin;
Original_name
----------------------
First
Second
SQL> purge recyclebin;

Testing:
SQL> select original_name from recyclebin;
MUARLI SIR 2015
16
No rows selected

4) Truncate: Oracle 7.0 introduced truncate command , whenever we are using “truncate”
command total data permanently deleted from table.
Syntax: truncate table tablename;
Example: create table first as select * from emp;
SQL> select * from first;
SQL> truncate table first;
Testing: SQL> select * from first;
No rows selected.
SQL> desc first;

5) Rename: It is used to rename a table and renaming a column also.

Renaming a Table:
Syntax: rename oldtablename to new tablename;
Example: rename first to last;

Renaming a Column: (oracle 9i)

Syntax: alter table tablename rename column oldcolumnname to newcolumnname;
Example: SQL> alter table emp rename column empno to sno;
SQL> select * from emp;
Note: In all database systems by default all DDL commands are automatically committed.

2. DATA MANIPULATION LANGUAGE (DML):

 Insert
 Update
 Delete
 Merge (oracle 9i)
These commands are used to manipulate data in a table.
1) Insert: It is used to insert data into a table.
Method1: Insert into tablename values(value1, value2, value3…..);
Example: SQL> create table first(sno number(10), name varchar(20));

Inserting data into table:

SQL> insert into first values(1, „abc‟);
SQL> insert into first values(2, „sachin‟);
SQL> select * from first;
Sno name
----------- -------------
1 abc
2 sachin
Method2: (using substitutional operator “&”)
Syntax: insert into tablename values(&col1, &col2, &col3……);
Example: SQL> insert into first values(&sno, „&name‟);
MUARLI SIR 2015
17
Enter value for sno: 3
Enter value for name: xyz
SQL> /
Enter value for sno: 4
Enter value for name: sachin
SQL> select * from first;

Date: 11/3/15

Method3: (skipping columns)

Syntax: insert into tablename(columnname1, columnname2,…..)values(value1,value2,…);
Example: SQL> insert into first(name)values(„www‟);
SQL>select * from first;

2) Update: It is used to change data within a table.

Syntax: update tablename set columnname=newvalue where columnname=oldvalue;
Example: SQL> update emp set sal=1000 where ename=‟SMITH‟;
SQL> select * from first;
Note: In all database if we want to insert data into particular cell then we must use update
command.
Example: SQL> alter table first add address varchar2(10);
SQL> update first set address=‟mumbai‟ where name=‟sachin‟;
SQL> select * from first;

Output: sno name address

----------------------------------------
1 abc
2 sachin Mumbai
3 xyz
4 zzz
Example: SQL> update first set address=null where name=‟sachin‟;
SQL> select * from first;

3) Delete: It is used to delete particular rows or all rows from a table;

Syntax:
 Delete from tablename;
 Delete from tablename where condition;
First command is to delete all rows
Second command is to delete particular rows.
Example: SQL> delete from first;
Rows deleted.
SQL> rollback;
SQL> select * from first;

MUARLI SIR 2015

18
Difference between „delete‟ and „truncate‟:
Whenever we are using „delete from tablename‟ or „truncate table tablename‟ then
automatically total data is deleted. Whenever we are using „delete from tablename‟ then
deleted data automatically stored in a buffer. We can get it back, this deleted data using
“rollback”. Whenever we are using „truncate table tablename‟ then total data permanently
deleted. Because truncate is an DDL command. That‟s why we cannot get it back this data
using “rollback”.

3) DATA RETRIVAL LANGUAGE(DRL) (OR) DATA QUERY LANGUAGE(DQL):

-> Select
In relational databases we can retrieve data fromtable using select statement.

SynSyntax: select col1,col2,…. From tablename where condition

group by columnname
having condition
order by columnname[asc/desc];

1. Select all cols and all rows.

2. Select all cols and particular rows.
3. Select particular cols and all rows.
4. Select particular cols and particular rows.

Date: 12/3/15
Creating a new table from another table:
Syntax: create table newtablename as select * from existingtablename;
Example: SQL> create table test as select * from emp;
SQL> select * from test;
Note: In all database systems whenever we are copying a table from another table
constraints are never copied. (Primary table, Foreign key,…..);
Creating a new table from existing table without copying data:
Syntax: create table newtablename as select * from existingtablename where
falsecondition;
Example: SQL> create table test1 as select * from emp where 1=2;
Testing: SQL> select * from test1;
No rows selected.
SQL> desc test1;
Operators Used in “Select” statement:

1. Arthematic operator(+,-,*,/)
2. Relational operator(=,<,<=,>,>=,[<> or!=]not equal).
3. Logical operator (AND,OR,NOT)
MUARLI SIR 2015
19
4. Special operator.

Arithmetic operator is used for “select” (eg: select column1, column2,…)

Relational and Logical and Special operators are used for “where” (eg: where condition)
We can also use arithmetic operators in “where” conditions.
 Arithmetic operators are used in number, data datatype column.
Q) write a query to display ename, sal, annsal from emp table;
Ans: SQL> select ename, sal, sal*12 annsal from emp;

ENAME SAL ANNSAL

----------- ----------- --------------------
SMITH 700 8400

Q) write a query to display the employees except job as clerk from emp table.
Ans: select * from emp where job<>‟CLERK‟; {or job != „CLERK‟}

Q) Write a query to display the employees who are getting more than 2000 salary from emp
table?
Ans: SQL> select * from emp where sal>2000;

Example: SQL> select * from emp where job=‟CLERK‟ and sal>2000;

Note: In databases if we want to retrieve multiple values within a single column then we
must use “OR” operator.
Example: SQL> select * from emp where job=‟CLERK‟ or job=‟SALESMAN‟;

Q) write a query to display the employees who are belongs to the department numbers
20,50,70,90?
Ans: SQL> select * from emp where deptno=20 or deptno=50 or deptno=70 or deptno=90;

 SQL> select * from emp where sal>2000 and sal<5000;

** Special operators:
1) „In‟ opposite „not in‟
2) „between‟ opposite „not between‟
3) „is null‟ opposite „is not null‟
4) „like‟ opposite „not like‟

1. In: It is used to pick the values one by one from list of values. „In‟ operator performance
is very high compared to „OR‟ operator. If we want to retrieve multiple values in a single
column then we are using „IN‟ operators in place of „OR‟ operator because „IN‟ operator
performance is high.
Syntax: select * from tablename where columnname in(list of values);
Columnname belongs to any data type is possible.
Example: SQL> select * from emp where deptno in(20,30,50,70,90);
MUARLI SIR 2015
20
Date: 13/3/15
Eg: SQL> select * from emp where ename in(„SMITH‟,‟FORD);
Note: In all datebase systems whenever we are using multiple row sub queries then we
must use “in” operator.
Eg: select * from emp where deptno not in(10,20,null);
Note: In all relational databases “not in” operator doesn‟t work with “null” values.
Null: Null is an undefined, unavailable, unknown value. It is not same as “zero”. Any
arithmetic operations performed on null values again it becomes “null”.
Eg: null+50=> null.
Q) Write a query to display ename, sal, comm, sal+comm of the employee SMITH from emp
table.
Ans: SQL> select ename, sal, comm, sal+comm from emp where ename=”SMITH”;

o/p: ename sal comm. Sal+comm

---------- ------- ----------- ----------------
SMITH 1100 null null
To overcome this problem oracle provided “NVL()” function.
NVL(): NVL() is a predefined function which is used to replace or substitute user defined
value in place of “null”.
Syntax: NVL(exp1,exp2);
Here exp1,exp2 must belongs to same datatype. If exp1 is null then it returns exp2.
Otherwise it returns exp1.
Eg: 1) NVL(null,30) -> 30.
2) NVL(10,20) -> 10.
Solution: select ename, sal, comm, sal+NVL(comm, 0) from emp where ename= „SMITH”;

ENAME SAL COMM SAL+COMM

--------- -------- ----------- ------------------
SMITH 1100 Null 1100

 Sal+ nvl(comm.,0)
 1100 + nvl(1100,0)
 1100 + 0
 1100

NVL2(): Oracle 9i introduced NVL2() function. This function accepts three parameters.
Syntax: NVL2(exp1,exp2,exp3).
Here if exp1 is null, then it returns exp3. Otherwise it returns exp2.
Eg: SQL> select nvl2(null,10,20) from dual;
20
SQL> select nvl2(30,40,50) from dual;
40
Q) Update employee commission as follows using nvl2() from emp table;
MUARLI SIR 2015
21
1) If comm is null then update comm.->500
2) If comm is not null then update comm-> 500
Ans: SQL> update emp set comm=nvl2(comm, comm+500,500);
SQL> update emp set comm=nvl2(comm,comm+500,500);
SQL> select * from emp;

2) Between: This operator is used to retrieve range of values. This operator is also call as
Betweem….. And operator.
Syntax: Select * from tablename where columnname between lowvalue and highvalue.
Eg: select * from emp where sal between 2000 and 5000;

Date: 14/3/15

While using “DATE”:

SQL> select * from emp where hiredate between ‟01-jan-1981‟ and ‟01-jan-1982‟;

3) Is null, is not null: These two operators used in „where” condition only. These two
operators are used to test weather a column having null values or not.
Syntax: select * from tablename where columnname is null;
Syntax: select * from tablename where columnname is not null;
Q) write a query to display the employees who are not getting commission from emp table.
Ans: SQL> select * from emp where comm is null;

Q) write a query to display the employees who are getting commission from emp table?
Ans: SQL> select * from emp where comm is not null;

4) Like: It is used to search strings based on character pattern. “Like” operator

performance is very high compare to searching functions. In all databases along with like
operator we are using 2 wild card charcters.

% -> group of characters if want to match or string

_ -> single character to match.

Syntax: select * from tablename where columnname like „characterpattern‟;

Q) write a query to display the employees whose ename start with “M” from emp table using
like operator.

Ans: SQL> select * from emp where ename like „M‟;

MUARLI SIR 2015
22
 MARTIN
MILLER
Q) write a query to display the employees where ename second letter will be „L‟ from emp
table using „like‟ operator?
Ans: SQL> select * from emp where ename like „_ _L%‟;
ALLEN
CLARK
BLAKE
Q) write a query to display the employees who are joining in the month December from the
emp table using like operator.
Ans: SQL> select * from emp where hiredate like „_ _ _D%‟; (or) „%DEC%”;

Q) write a query to display the employees who are joining in the year “81” from emp table
using like operator?
Ans: SQL> select * from emp where emp table like „%81‟;

“Escape” function used in like operator:

Whenever wild card characters available in character data then we are trying to retrieve the
data using like operator then database servers treated these wild card characters as special
characters to overcome this problem ansi/iso SQL provided a special function “Escape” along
with „like‟ operator. Which is used to escape special characters and also this function treated
as _,% as same meaning as _,%.
Syntax: select * from tablename where columnname like „character pattern‟ escape „escape
character‟;
Note: Escape character must be an single character having length „1‟ within character
pattern we must use escape character before wild card character.
Eg: SQL> insert into emp(empno, ename) values(1,‟S_ MITH‟);
SQL> select * from emp;
Q) write a query to display the employees whose ename start with „S_‟ from emp table using
„like‟ operator?
Ans: SQL> select * from emp where ename like „S_%‟;
SMITH
SCOTT
S_MITH
Solution: SQL> select * from emp where ename like „S?_%‟ escape‟?‟;
S_MITH
Eg: insert into emp(empno, ename) values(2,‟S__MITH‟);
SQL> select * from emp;
Solution: SQL> select * from emp where ename like „S?_?_%‟ escape‟?‟;
S__MITH

 Concatenation Operator(|| double pipe): It is not a special operator rarely used in

“SQL” and regularly used in “PL/SQL”. If we want to display column data along with literal
strings then we must use concatenation operator.
MUARLI SIR 2015
23
 Eg: select „my employee names are‟|| ename from emp;
If we want to display our own space in between the columns then also we can use
concatenation operator.
Eg: select ename||‟ „||sal from emp;

Date: 16/3/15
Functions: Functions are used to solve particular task and also functions must return a
value. Oracle have two types of functions.
1. Predefined functions
2. User defined functions
1) Predefined functions: there are 4 types.
 Number function
 Character function
 Date function
 Group function(or) Aggregate function
1) Number function: These functions operate over “number” data.

 Abs() : It is used to convert negative values into positive values.

Eg: SQL> select abs(-50) from dual;
50
 Dual(): Dual is a predefined virtual table which contains only one row and one column
Dual table is used to test predefined, user defined functions functionality.
Example for testing predefined functions:
SQL> select nvl(null,30) from dual;
30
SQL> select nvl(20,30) from dual;
20
SQL> select * from dual;

Note: In oracle by default dual table column datatype is varchar2().

Eg: select * from dual;

Generally, this table is also used to perform mathematical operations.

Eg: SQL> select 10+50 from dual;
60
Eg: SQL> select ename, comm, sal, abs(comm-sal) from emp where comm is not null;

 Mod(m,n): It will gives remainder after „m‟ divided by „n‟.

Eg: SQL> select mod(10,5) from dual;

0

MUARLI SIR 2015

24
 Round(m,n): It rounds given floated valued number „m‟ based on „n‟.
Eg: SQL> select round(1.8) from dual;
2
SQL> select round(1.23456,3) from dual;
1.235
Note: Round always checks remaining number if remaining number is above 50% then
one added to the rounded number.
Eg: SQL> select round(1285.456,-1) from dual;
1290
Execution:
Step1: 1280->5 is replace with 0. Out of 10, 5 is above 50%(>=)
Step2: 1280
+1
---------------
1290
Eg: select ename, sal, round(sal,-1) from emp;
Eg: select ename, sal, round(sal/22), round(sal/22,1), round(sal/22,2) from emp;
 Trunc(m,n): It truncates given floated values number „m‟ based on „n‟. This function
doesn‟t check remaining number is above 50% or below 50%.
Eg: SQL>select trunk(1.8) from dual;
Eg: SQL> select trunk(1.23456,3) from dual;
1.234
 Greatest(exp1,exp2,…. expn), Least(exp1,exp2,…. expn):
Greatest returns maximum value among given expressions. Where as Least returns
minimum value among given expressions.
Eg: SQL> select greatest(3,5,8,9) from dual ;
9
SQL> select ename, sal, comm, greatest(sal,comm)from emp where comm is not null;
SQL> select max(sal) from emp;
6600

Date : 18/3/15

 Ceil() and Floor(): These two functions always return integer. Ceil() returns nearest
greatest integer where as floor() returns nearest lowest integer.
Eg: select ceil(1.3) from dual;
2
Eg: select floor(1.9) from dual;
1
2) CHARCTER FUNCTIONS:
 Upper(): It is used to convert a string or column values into “ upper case “.
Eg: select upper(„abc‟) from dual;
ABC
For column: SQL> select upper(ename) from dual;
MUARLI SIR 2015
25
 Lower():
Eg: select lower(ename) from emp;
Updating or modifying data within table:
SQL> update emp set ename=lower(ename);
 Initcap(): It returns first letter is capital and all remaining letters are small.
Eg: SQL> select initcap(ename) from emp;
Eg: SQL> select initcap(„ab cd ef‟) from emp;
Ab Cd Ef
 Length(): It returns number datatype that is ., it returns total length of the string
including spaces.
Eg: select length(„AB_CD‟) from dual;
5
 Substr(): It will extract portion of the string within given string based on last two
parameters.
Eg: select substr(„ABCDEF‟,2,3) from dual;
BCD
SQL> select substr(„ABCDEF‟,-2,3) from dual;
EF
SQL> select substr(„ABCDEF‟,-4) from dual;
CDEF
Syntax: substr(columnname (or) „string name‟, starting position, no. of characters position);
Starting position-> can be +ve or –ve
Starting position and no. of characters position-> must be numbers

Q) write a query to display the employees whose ename second letter would be “LA” from
emp table using substring function?
And: SQL> select * from emp where substr(ename,2,2)=‟LA‟;
BLAKE
CLARK
Note: In all database systems we are not allowed to use group functions in where clause but
we are allowed to use number functions, character functions, date functions in where clause.
Eg: select * from emp where sal=max(sal);
Error: group function is not allowed here.
Q) write a query to display the employee whose employees length is 5 from emp table.
Ans: SQL> select * from emp where length(ename)=5;

 Instr(): instr() always returns number datatype that is it returns position of the
delimiter, position of the character, position of the string within given string .
Eg: SQL> select instr(„ABC*D‟,‟*‟) from dual;

Date: 19/3/15
Eg: SQL> select instr(„ABCDEFGHCDKMNHCDJL‟,‟CD‟,-5,2) from dual;
3
SQL> select instr(„ABCDEFGHCDKMNHCDJL‟,‟CD‟,-4,2) from dual;
MUARLI SIR 2015
26
 9
Syntax: instr(columnname(or)‟string name‟, „str‟, searching position, No. of occurrences from
searching position);
Searching position -> +ve or –ve
Searching and No. of occurrences -> must be numbers.
Always in string returns position based on last two parameters but oracle server counts no. of
characters left side first position onwards.

Lpad(): It will fills remaining spaces with specified character on the left side of the given
string. Here always second parameter returns total length of the string. And also third
parameter is an optional.Syntax: Lpad(columnname (or) „string name‟, total length, „filled
character‟);
Total length -> number

Eg: SQL> select lpad(„ABCD‟,10,‟#‟) from dual;

######ABCD
SQL> select lpad(„ABC‟,5) from dual;
(space)(space)ABC
SQL> select lpad(„ABC‟,2,‟*‟) from dual;
AB

Rpad():
Eg: SQL> select rpad(„ABCD‟,10,‟#‟)from dual;
ABCD######
SQL> select rpad(ENAME,50,‟-„)||sal from emp;

Ltrim: It removies specified character on the left side of the given string.
Syntax: Ltrim(columnname(or)‟string name‟, {set of characters});

Eg: SQL> select ltrim(„SSMISSTHSS‟,‟S‟)from dual;

MISSTHSS
SQL> select job,ltrim(job,‟CSM‟) from emp;
JOB LTRIM(JOB)
----------------- -------------------------
CLERK LERK
SALESMAN ALESMAN
MANAGER ANAGER

Rtrim:
Eg: SQL> select rtrim(„SSMISSTHSS‟,‟S‟) from dual;
SSMISSTH
Trim(): Oracle 8i introduced trim function it is used to remove left and right side specified
characters and also it is used to remove leading and trailing spaces.
Syntax: trim(„character‟ from „string name‟);
MUARLI SIR 2015
27
Eg: SQL> select trim(„S‟ from „SSMISSTHSS‟)from dual;
MISSTH
This function also behaves like a ltrim, rtrim based on leading and trailing claused.
Eg: select trim(leading „s‟ from „SSMISSTHSS‟) from dual;
SSMISSTH
Eg: select length(trim(„ welcome „))from dual;
7

Translate() and Replace():

Translate() is used to replaces character by character where as replace() is used to replaces
character by string (or) string by string.

Eg: select translate(„india‟,‟in‟,‟xy‟),replace(„india‟,‟in‟,‟xy‟) from dual;

Translate Replace
Xydxa xydia

Syntax: translate(„str1‟,‟str2‟,‟str3‟….);
Eg: select translate(„ABCDEF‟,‟FEDCBA‟,123456);
654321
Eg: select replace(„ABC‟,‟ „,‟india‟)from dual;
AindiaBindiaC
Eg: select job,replace(job, „SALESMAN‟,‟MARKETING‟) from emp;
Note: In replace function if we are not specifying third parameter specified character
permanently removed from the string.
Eg: select replace(„SSMISSTHSS‟,‟S‟) from dual;
MITH

Date: 20/3/15
If you want to count number of times particular character occurs within a given string then
also we are using replace function along with length function.
Q) Write a query to count number of times that particular „I‟ occurred within given string
„india‟ using replace function?
Ans: SQL> select length(„india‟)-length(replace(„india‟,‟I‟)) from dual;
2
Concat(): It is used to concatenate given two strings.
Eg: select concat(„wel‟,‟come‟)from dual;
Welcome

DATE functions: In oracle by default date format is DD-MON-YY. Oracle having following
date functions.
1. Sysdate
2. Add_months()
3. Last_day()
MUARLI SIR 2015
28
4. Next_day()
5. Months_between()

1) Sysdate: It returns current date of the system in the oracle date format
SQL> select sysdate from dual;
20-MAR-15
2) Add_months(): It is used to add or subtract number of months to the specified date
based on second parameter.
Syntax: add_months(date, number);
Number-> can be positive or negative.
Eg: SQL> select add_months(sysdate,1)from dual;
20-Apr-15
SQL> select add_months(sysdate,-1) from dual;
20-FEB-15
3) Last_day(): It returns last date of the given months based on the passed date. This
function always accepts one parameter.
Syntax: last_day(date);
Eg: SQL> select last_day(sysdate)from dual;
31-Mar-15
4) Next_day(): It returns next occurrence day from the specified date based on second
parameter.
Syntax: next_day(date,‟day‟);
Eg: select next_day(sysdate,‟monday‟)from dual;
23-MAR-15
5) Months_between(): These function always returns “number” data type. i.e., it returns
number of months between two specified dates.
Syntax: months_between(date1,date2)
Here date1 is more than date2 otherwise this function returns “negative” sign.
Eg: SQL> select ename, round(months_between(sysdate,hiredate))from emp;

DATE arithmetic:
1. Date + number
2. Date – number
3. Date1 + date2
4. Date1 – date2
Eg: SQL> select sysdate+1 from dual;
SQL> select sysdate-1 from dual;
SQL> select sysdate-sysdate from dual;
0
Q) write a query to display first date of the current month using predefined date functions
sysdate, add_months(), last_date() functions.
Ans: SQL> select last_day((add_months(sysdate,-1)))+1 from dual;
OR
SQL> select add_months(last_day(sysdate),-1)+1 from dual;
MUARLI SIR 2015
29
01-MAR-15

DATE CONVERSION FUNCTIONS:

Oracle provided two date conversion function. They are:
1. To_char()
2. To_date()

Date: 23/3/15

1. to_char(): It is used to convert date time into character type i.e., it converts date type
into date string.
Eg: SQL> select to_char(sysdate,‟dd/mm/yy‟)from dual;
23/03/15
Note: Basically “to_char” character formats are case sensitive formats.
Eg: SQL> select to_char(sysdate,‟DAY‟)from dual;
MONDAY
Day-> Monday(lowercase letters)
DAY
DY
MONTH
MON
YEAR
RETURNS CHARACTER AS OUTPUT
D
DD
DDD
MM
YY
YYYY
HH
MI
SS
RETURNS NUMBER AS OUTPUT
Eg: SQL> select to_char(sysdate,‟DY‟)from dual;
MON

Eg: SQL> select to_char(sysdate,‟D‟)from dual;

2
D->day of the week(sun-1,mon-2,tue-3,….sat-7)
DD->day of the month
DDD->day of the year(tells how many days completed)
DDTH-> used as date with „rd‟ or „th‟ (eg: 5th 19th 23rd)
Eg: SQL> select to_char(sysdate,‟DDSPTH‟) from dual;
Twenty third
MUARLI SIR 2015
30
SP-> spell out
Eg: SQL> select to_char(sysdate,‟MON‟) from dual;
MAR
Eg: SQL> select to_char(sysdate,‟HH:MI:SS‟) from dual;
09:51:42
By default 12-hours format
to convert into 24-hours format we have to “HH24:MI:SS” (24-hours format)

Q) SQL> select to_char(‟15-JUN-05‟,‟15-JUNE-05‟) from dual;

Error: invalid number
Note: whenever we are using to_char() always first parameter must be oracle “date” type
otherwise oracle server returns an error.

2. to_date(): It is used to convert date string into oracle “date” type(oracle date format)
Eg: SQL> select to_date(‟27/JUNE/05‟) from dual;
27-JUN-05
Eg: SQL> select to_date(„27/06/05‟) from dual;
Error: invalid month (or) not a valid month
Whenever are using to_date() passed parameter return values match with the default date
format return values. Otherwise oracle server returns an error. To overcome this problem use
a second parameter as same as first parameter format then only oracle server automatically
converts “date string” into “date type”.
Solution: SQL> select to_date(‟27/06/05‟,‟DD/MM/YY‟) from dual;
27-JUN-05

 SQL> select to_date(‟09-feb-15‟)+5 from dual;

14-feb-15
 SQL> select to_date(‟09-feb-15‟,‟DD-MM-YY‟)+5 from dual;
14-FEB-15
Q) Write a query to convert given date string into client requirement format using to_char()
Given date is -> ‟15-jun-05
Expected format-> „15/jun/05‟
SQL> select to_char(‟15-jun-05‟,‟dd/month/yy‟) from dual;
Error: invalid number
SQL> select to_char(to_date(‟15-jun-05‟),‟dd/month/yy‟) from dual;
15/jun /05
Date: 24/3/15

Fill mode(FM): Whenever we are using to_char() if format is either month(or) day then
oracle server automatically returns “spaces” when the month (or) day value less than the
maximum length of the month (or) day. To overcome this problem oracle provided Fill
Mode(FM) which suppress blank spaces and leading zeros(0). This mode is used in second
parameter of the to_char().
Eg: SQL> select to_char(to_date(‟15-JUN-05‟),‟DD-MON-YY‟) from dual;
MUARLI SIR 2015
31
15-JUNE -05
SQL> select to_char(to_date(‟15-JUN-05‟), „DD/FMMONTH/YY‟) from dual;
15/JUNE/05
Q) Write a query to display the employees who are joining in the month December from
“emp” table using to_char()?
Ans: SQL> select * from emp where to_char(HIREDATE,‟MON‟)= „DEC‟;
OR
SQL> select * from emp where to_char(HIREDATE,‟MM‟)=‟12‟;
OR
SQL> select * from emp where to_char(HIREDATE,‟FMMONTH‟)=‟DECEMBER‟;

Q) Write a query to display the employees who are joining in the year „81‟ from emp table
using to_char()?
Ans: SQL> select * from emp where to_char(hiredate,‟yy‟)=‟81‟;
SQL> select hiredate, to_char(hiredate,‟yyyy‟) from emp;
In Oracle, whenever we are passing date string into date functions then oracle server
automatically converts date string into date type.
Eg: SQL> select last_day(‟12-JUN-05‟) from dual;
30-JUN-05
But here passed parameter format must be default date format otherwise oracle server
returns an error. To overcome this problem we must use to_date().
Eg: SQL> select last_day(„12-06-05‟) from dual;
Error: invalid month
Solution: SQL> select last_day(to_date(‟12-06-05‟,‟DD-MM-YY‟)from dual;
30-JUN-05
In Oracle whenever we are inserting dates into date data type column then oracle server
automatically converts date string into date type when inserted date string is in oracle date
format. Otherwise oracle server returns an error. To overcome this problem also then we must
use to_date().
Eg: SQL> create table test(col1 date);
SQL> insert into test values(sysdate);
SQL> select * from test;
COL1
------------
24-MAR-15
SQL> insert into test values(‟15-aug-05‟);
SQL> select * from test;
COL1
------------
24-MAR-15
15-AUG-05
SQL> insert into test values(‟15-08-05‟);
Error: not a valid month
SQL> insert into test values(to_date(‟15-08-05‟,‟DD-MM-YY‟));
MUARLI SIR 2015
32
1 row inserted
SQL> select * from test;
COL1
------------
24-MAR-15
15-AUG-05
15-AUG-05

Round(), Trunc() used in dates:

In Oracle, “date” data type contains both date and time. Whenever we are using the round()
or trunc() then automatically date part changed based on time portion.

Date: 25/3/15

Round() adds one day to the given date if time portion is greater than or equal to 12 noon
and also time portion automatically set to “zero”.
Eg: SQL> select to_char(sysdate,‟DD-MON-YYYY HH24:MI:SS‟) from dual;
25-MAR-2015 12:24:59
When applying round():
SQL> select to_char(round(sysdate),‟DD-MON-YYYY HH24:MI:SS‟) from dual;
26-MAR-2015 00:00:00
Whenever we are using trunc() oracle server automatically returns same date if time portion
is greater than or equal to 12 noon also and also trunc() time portion is set to “zero”(0)
Eg: SQL> select to_char(trunc(sysdate),‟DD-MON-YYYY HH24:MI:SS‟) from dual;
25-MAR-2015 00:00:00
Q) Write a query to display the employees who are joining today from emp table?
Ans: SQL> select insert into emp(empno,ename,hiredate)values(1, „MURALI‟,25-MAR-15);
SQL> select * from emp;
SQL> select * from emp where trunc(hiredate) = trunc(sysdate);

GROUP FUNCTION (OR) AGGREGATE FUNCTION:

Oracle having following group function
1. max()
2. min()
3. avg()
4. sum()
5. count(*)
6. count(column name)
In all databases group functions are operate over number of values within a column and
returns a single value
1. Max(): It returns maximum value within a column.
Eg: SQL> select max(sal) from emp;
6600
SQL> select max(hiredate) from emp;
MUARLI SIR 2015
33
23-MAY-87
SQL> select max(ename) from emp;
WARD (A=1……… Z=26)

2. Min():
Eg: SQL> select min(sal) from emp;
750
To know the senior most employee
SQL> select min(hiredate) from emp;
17-DEC-80
In all database systems we are not allowed to use group functions in “where” clause.
SQL> select * from emp where sal= min(sal);
Error: group function is not allowed in “where”

3. Avg(): It returns average from number data type column.

Eg: SQL> select avg(sal) from emp;
2473.21429
SQL> select avg(comm) from emp;
550
NOTE: In all database systems by default all group function ignores null values except
“count(*)”.
To count all null values also:
SQL> select avg(nvl(comm,0)) from emp;
157.142857

4. Sum(): It returns total from number data type column.

Eg: SQL> select sum(sal) from emp;
34625

5. Count(*): It counts number of rows in a table(including null values)

Eg: SQL> select count(*) from emp;
6. Count(column name): It counts number of not null values within a column.
SQL> select count(comm) from emp;
4
SQL> select count(distinct(dept no)) from emp;
3
GROUP BY: This clause is used to divide similar data items into set of logical groups.
Syntax: select columnname….. from tablename group by columnname;

Data: 26/3/15
Q) Write a query number of employees in each department from emp table using “group by”?
Ans: SQL> select deptno,count(*) from emp group by deptno;
DEPTNO COUNT(*)
------- ----------
MUARLI SIR 2015
34
 30 6
20 5
10 3
Q) Write a query to display number of employees in each job from emp table using group by?
Ans: SQL> select job, count(*) from emp group by job;
JOB COUNT(*)
--------- ----------
CLERK 4
SALESMAN 4
PRESIDENT 1
MANAGER 3
ANALYST 2
Q) Write a query to display deptno, minimum and maximum salary from emp using group by?
Ans: SQL> select deptno, min(sal), max(sal) from emp group by deptno;
DEPTNO MIN(SAL) MAX(SAL)
------- ---------- ----------
30 950 2850
20 800 3000
10 1300 5000

Note: In all database systems we can also use “group by” clauses without using “group”
functions.
Eg: SQL> select deptno from emp group by deptno;
RULE: Other than group function columns specified after select those all columns must be
used in after “group by”. Otherwise oracle server returns an error “not a GROUP BY
expression”.
Eg: SQL> select deptno, max(sal), ename from emp group by deptno,ename;
ERROR at line 1:
ORA-00979: not a GROUP BY expression

SQL> select deptno from emp group by deptno,job;

“Reverse also possible”.
Note: Whenever we are submitting “group by” clauses then database servers select the data
from the table from based on the group by clause columns. These columns space database
servers group the data in the result set, from this result only we are selecting the data using
number of columns after select keyword.
Note: Whenever we are using group functions without using “group by” clause then database
servers executes these group functions based on all values in a column. Whereas when we are
using “group by” then these group functions executed each and every group wise in a table.
Note: Generally in all database systems we are not allowed to display group function with
another columns to overcome this problem we must use “group by” clause.
Eg: step1: SQL> select sum(sal) from emp;
34627
Step2: SQL> select deptno, sum(sal) from emp;
Error: not a single-group group function

MUARLI SIR 2015

35
Solution: SQL> select deptno, sum(sal) from emp group by deptno;
DEPTNO SUM(SAL)
------- ----------
30 9400
20 10875
10 8750
Eg: SQL> select deptno, job, sum(sal), count(*) from emp group by deptno, job;
DEPTNO JOB SUM(SAL) COUNT(*)
------- --------- ---------- ----------
20 CLERK 1900 2
30 SALESMAN 5600 4
20 MANAGER 2975 1
30 CLERK 950 1
10 PRESIDENT 5000 1
30 MANAGER 2850 1
10 CLERK 1300 1
10 MANAGER 2450 1
20 ANALYST 6000 2

Date: 27/3/15

Q) Write a query to display those departments having more than 3 employees from emp
table?
Ans: SQL> select deptno,count(*) from emp group by deptno where count(*)>3;
Error: SQL commond not properly ended.
Solution: SQL> select deptno, coutn(*) from emp group by deptno having count(*)>3;

HAVING Clause: In all databse systems after “group by” clause we are not allowed to use
“where” clause. In place of this one ansi/iso SQL provided another clause “having”. Generally
if we want to restrict groups after “group by” then we must use “having” clause. Generally in
“where” clause we are not allowed to use group functions where as in having clause we can
also use group functions.
Q) Write a query to display those departments having more than 10,000 sun(sal) from emp
table?
And: SQL> select deptno, sum(sal) from emp group by deptno having sum(sal)>10,000;

Deptno sum(sal)
-----------------------
10 13550
20 12675

Q) Write a query to display year, number of employees per year in which more than one
employee was hired from emp table using “group by”?
And: SQL> select to_char(hiredate,‟YYYY‟), count(*) from emp group
by(to_char(hiredate,‟YYYY‟));

MUARLI SIR 2015

36
 OR
SQL> select to_char(hiredate,‟YYYY‟) year, count(*) from emp
group by to_char(hiredate,‟YYYY‟)
having count(*)>1;

year count(*)
------------------------
1981 10
1982 2

Invisible columns:
Eg: SQL> select deptno, sum(sal) from emp
group by deptno
having sum(sal)>10000;
Here count(*) is invisible column.
SQL> select deptno, sum(sal) from emp
Group by deptno
Having count(*)>3;
Deptno sum(sal)
---------------------------
20 12675
30 8400
ORDER BY: This clause is used to arrange the data in sorting order along with “order by”
clause we are using two keywords.

ORDER BY

asc(ascending desc(descending)

By default
) „order by‟ clause having “Ascending order”.
Syntax: SQL> select * from tablename order by columnname[asc/desc];
Eg: SQL> select sal from emp order by sal;
SQL> select deptno, sal from emp
order by deptno, sal;
Eg: SQL> select deptno, count(*) from emp
Where sal>1000
group by deptno
having count(*)>3
order by depnto desc;

Deptno count(*)
----------------------
30 4
MUARLI SIR 2015
37
20 5

Date: 28/3/15

Note: In oracle we can also use column position in place of columnname within “order by”
clause which is used to improve performance query.
Eg: SQL> select * from emp order by 6 desc;

ROLLUP, Cube:
Oracle 8i introduced rollup, cube clauses. These clauses are used along with group by clause
only. These clauses are used to calculate subtotal, grand total automatically.
Syntax: select col1, col2,…….. from tablename
group by rollup(col1,col2,..);
Syntax: select col1, col2,…. From tablename group by cube(col1,col2,…);
Rollup is used to calculate subtotal value based on a single column where as if we want to
calculate subtotal value based on number of column wise then we must use cube.
Eg: SQL> select deptno, job, sum(sal)from emp group by rollup(deptno, job);
SQL> select deptno, job, sum(sal), count(*) from emp group by cube(deptno,job);
Eg: SQL> select ename, sum(sal) from emp group by rollup(ename);

JOINS: Joins are used to retrieved data from multiple tables. In all databases when we are
joingin “n” tables then we must use “n-1” joining conditions. Oracle having following types of
joins.
1. Equi joing (or) Inner join
2. Non equi join
3. Self join
4. Outer join

These 4 joins are also called as “8i joins”.

9i joins or ansi joins.
1. Inner join
2. Left outer join
3. Right outer join
4. Full outer join
5. Natural join

In oracle, we can also retrieve data from multiple tables without using join condition in this
case oracle server internally uses cross join. But cross join is implemented based on Cartesian
product that‟s why this internal join returns more duplicate data.
Eg: SQL> select ename, sal, dname, loc from emp, dept;

1. Equi join (or) Inner join: Based on equality condition we are retrieving data from
multiple tables. Here joining conditional columns must belongs to same datatype.
Generally, when tables having common columns then only we are using this “join”.
MUARLI SIR 2015
38
Syntax: select col1, col2,….
from tablename1, tablename2
where tablename1.commoncolumnname= tablename2. Commoncolumnname;

tablename. Commoncolumnname is a join condition

Eg: SQL> select ename, sal, deptno, dname, loc from emp, dept where emp.deptno=
dept.deptno;
Error: column ambiguously defined.
Solution: SQL> select ename, sal, dept.deptno, dname, loc from emp, dept where
emp.deptno = dept.deptno;

Date: 30/3/15

NOTE: Generally, to avoid ambiguity in future we must specify every “column name” along
with table name using „.‟ (Full stop or dot) operator.
NOTE: In all databases we can also create alias names for the table in “from” clause of the
“join conditions”. These alias names are also called as “Reference names”. These reference
names are used in select list, used in joining condition.
Syntax: from tablename aliasname1, tablename aliasname2.
These alias names must be different names.
USING ALIAS NAMES:
SQL> select ename, sal, d.deptno,dname, loc from emp e, dept d
Where e.deptno = d.deptno;

ENAME SAL DEPTNO DNAME LOC

---------- ---------- ---------- -------------- ----------
CLARK 2450 10 ACCOUNTING NEW YORK
KING 5000 10 ACCOUNTING NEW YORK
MILLER 1300 10 ACCOUNTING NEW YORK
JONES 2975 20 RESEARCH DALLAS
FORD 3000 20 RESEARCH DALLAS
ADAMS 1100 20 RESEARCH DALLAS
SMITH 800 20 RESEARCH DALLAS
SCOTT 3000 20 RESEARCH DALLAS
WARD 1250 30 SALES CHICAGO
TURNER 1500 30 SALES CHICAGO
ALLEN 1600 30 SALES CHICAGO
JAMES 950 30 SALES CHICAGO
BLAKE 2850 30 SALES CHICAGO
MARTIN 1250 30 SALES CHICAGO
In the above query , here deptno(“40”) doesn‟t displayed if we are using dept.deptno also.
NOTE: In all databases always equi joins returns matching rows only.
Q) Write a query to display the employees who are working in the location Chicago from emp,
dept tables using equi join?
Ans: SQL> select loc, ename from emp, dept where emp.deptno=dept.deptno group by loc
having loc=‟chicago‟;
error: not a GROUP BY expression
MUARLI SIR 2015
39
Solution: SQL> select ename, loc from emp, dept
where emp.deptno=dept.deptno
and loc=‟chicago‟;
ENAME LOC
---------- ---------
WARD CHICAGO
TURNER CHICAGO
ALLEN CHICAGO
JAMES CHICAGO
BLAKE CHICAGO
MARTIN CHICAGO
Note: If we want to filter the data after joining condition then we must use “and” operator in
8i joins where as in 9i joins we are using either “and” operator or “where” clause also.
Q) Write a query to display dname, sum(sal) from emp, dept tables using equi join?
Ans: SQL> select dname,sum(sal) from emp e,dept d where e.deptno=d.deptno group by
dname;
DNAME SUM(SAL)
-------------- ----------
ACCOUNTING 8750
RESEARCH 10875
SALES 9400

SQL> select dept.deptno, dname, sum(sal) from emp, dept where emp.deptno=dept.deptno
group by dept.deptno, dname;
DEPTNO DNAME SUM(SAL)
------- -------------- ----------
10 ACCOUNTING 8750
20 RESEARCH 10875
30 SALES 9400
Q) Write a query to display number of employees, min(sal), max(sal), sum(sal) in each
location from emp, dept tables using equi join?
Ans: select loc, count(*), min(sal), max(sal),sum(sal) from emp, dept where
emp.deptno=dept.deptno group by loc;
LOC COUNT(*) MIN(SAL) MAX(SAL) SUM(SAL)
------------- ---------- ---------- ---------- ----------
NEW YORK 3 1300 5000 8750
CHICAGO 6 950 2850 9400
DALLAS 5 800 3000 10875

SQL> select dname, sum(sal) from emp e, dept d where e.deptno=d.deptno group by dname
having dname;
DNAME SUM(SAL)
-------------- ----------
RESEARCH 10875
SQL> select dname, sum(sal) from emp e, dept d where e.deptno=d.deptno group by
rollup(dname);
DNAME SUM(SAL)
-------------- ----------
ACCOUNTING 8750
RESEARCH 10875
SALES 9400
29025

MUARLI SIR 2015

40
2. Non Equi Join: Based on other than equality condition(not equal to, >,<,>=,<=,between,
in,….) we can retrieve data from multiple tables. In oracle, this join is also called as
“Between….. and Join”.

Date: 31/3/15

Eg: SQL> create table test1(deptno number(10),ename varchar2(20));

SQL> insert into test1 values(10,‟a‟);
SQL> insert into test1 values(20,‟b‟);
SQL> select * from test1;
DEPTNO ENAME
---------------------
10 a
20 b
SQL> create table test2(deptno number(10),dname varchar2(20));
SQL>insert into test2 values(10,‟c‟);
SQL>insert into test2 values(20,‟d‟);
SQL>insert into test2 values(30,‟e‟);
SQL> select * from test2;
DEPTNO DNAME
----------------------
10 c
20 d
30 e
SQL> select * from test1,test2 where test1.deptno>test2.deptno;
DEPTNO ENAME DEPTNO DNAME
---------------------------------------------
20 b 10 c
NOTE: In oracle, we can also use non-equi join when tables doesnot have common columns.
Eg: select * from salgrade
SQL> select ename,sal,losal,hisal from emp,salgrade where sal between losal and hisal;
OR
Where sal>=losal and sal<=hisal;
3.SELF JOIN: Joining a table to itself is called “self join”. Here joining conditional columns
must belongs to same datatype. Before we are using self join we must create alias names for
the table in “from” clause.
Syntax: from tablename aliasname1,tablename aliasname2;
Generally, if we want to compare two different column values in a single table then we must
use self-join. But here these two columns must belong to same datatype.
Q) Write a query to display employee names and their manager names from emp table using
self join?
Ans: SQL> select e1.ename “employees”, e2.ename “managers” from emp e1, emp e2
Where e1.mgr=e2.empno;

MUARLI SIR 2015

41
SQL> select e1.ename “employees”, e1.mgr, e2.empno, e2.ename “managers” from emp e1,
emp e2
Where e1.mgr=e2.empno;
Q) Write a query to display the employees who are getting more salary than their manager
salary from emp table using “self join”?
Ans: SQL> select e1.name “employees”, e1.sal, e2.sal, e2.ename “manager”
from emp e1, emp e2
where e1.mgr=e2.empno
and e1.sal>e2.sal;
employees sal sal manager
----------------------------------------------
SCOTT 3600 3375 JONES
FORD 3600 3375 JONES

Q) Write a query to display the employees who are joining before their managers from emp
table using self join?
Ans: SQL> select e1.name “employees”,e1.hiredate, e2.hiredate,e2.ename “manager”
from emp e1,emp e2
where e1.mgr=e2.empno
and e1.hiredate< e2. Hiredate;

Date: 1/4/15
When to use Alias names:

In all databases systems whenever we are refer same table more than one time foe wuery
results then we must create “table alias” names. These alias names are also called as
„”Reference names”. These alias names internally behaves like a tables when query
execution time. These alias names must be different names.

Syntax: from tablename alias name1, tablename aliasname2;

Q) Write a query to display the employees who are getting same salary as scott salary
from emp table using table alias names?

Ans: SQL> select e2.ename,e1.sal from emp e1, emp e2 where e1.sal=e2.sal and
e1.ename=‟SCOTT‟;

4. OUTER JOIN: This join is used to retrieve all rows from one table and matching rows
from another table. Generally, using “equi join” we are retrieving matching rows only. If
we want to retrieve non-matching rows also then we are using join operator(+).

Within joining condition of “equi join” this join is called “ORACLE 8i Outer join”.

NOTE: Join operator can be used only one side at a time one joining condition.

MUARLI SIR 2015

42
Eg: SQL> select ename, sal, d.deptno,dname,loc from emp e, dept d where
e.deptno(+)=d.deptno;

e.deptno- matching rows

d.deptno- all rows

9i JOINS or ANSI JOINS:

1. Inner join
2. Left outer join
3. Right outer join
4. Full outer join
5. Natural join
1. Inner join: This join also returns matching rows only. Whenever tables having
common columns then only we are using this join. Here also joining conditional columns
must belongs to some data type. Inner join (9i) performance is very high compare to
oracle 8i equi join.
Eg: SQL> select ename, sal, d.deptno, dname, loc from emp e join dept d on e.
deptno=d.deptno;

Q) Write a query to display the employees who are working in the location „CHICAGO‟
from emp, dept tables using 9i inner join.
Ans: SQL> select ename, loc from emp e join dept d on e.deptno=d.deptno where
loc=‟CHICAGO‟;

Eg: SQL> create table t1(a varchar2(10), b varchar2(10),c varchar2(10));

SQL > insert into t1 values(„x‟,‟y‟,‟z‟);
SQL > insert into t1 values(„p‟,‟q‟,‟r‟);
SQL> select * from t1;
SQL> create table t2(a varchar2(10), b varchar2(10));
SQL>insert into t2 values(„x‟,‟y‟);
SQL.> select * from t2;

A B C A B
-------------------------- ---------------------
X Y Z X Y
P q r
8i equi join:
SQL> select * from t1,t2 where t1.a=t2.a and t1.b=t2.b;
9i inner join:
SQL> select * from t1 join t2 on t1.a=t2.a where t1.b=t2.b;
Output:
A B C A B

MUARLI SIR 2015

43
-------------------------- ---------------------
X Y Z X Y

Date: 2/4/15

Using clause:
In 9i joins we can also use “USING” clauses in place of “ON” clause. “Using” clause
performance is very high compare to “ON” clause and also when we are using “USING”
clause then oracle server automatically returning common columns on time only.
Syntax: select * from tablename1 join tablename2 using (common columnnames);
Eg: SQL> select * from t1 join t2
using(a,b);
Output:

A B C
-----------------
X y z
Note: whenever we are using “Using” clause then we are not allowed to use alias name or
joining conditional columns.
Eg: Select ename,sal,deptno,dname,loc
from emp e join dept d
using (deptno);

2. Left Outer Join:

This join always reuturns all rows from left side table and matching rows from reight side
table and also returns “NULL” values in place of non-matching rows in another table.
Eg: select * from t1;
A B C
-----------------
X y z
P q r
SQL> select * from t2;

A B
---------
X y
S t
SQL> select * from t1 left outer join t2 on t1.a=t2.a and t1.b=t2.b;
A B C A B
------------------- ---------------------
X Y Z X Y
P q r (null) (null)

MUARLI SIR 2015

44
3. Right Outer Join:
This Join returns all rows from right side table and matching rows from left side table and
also returns null values in place of “non-matching” rows in another table.

4. Full Outer Join:

This join returns all data from all the tables it is the combination of left and right outer
joins. This join always returns matching and non-matching rows. And also this join returns
null values in place of non-matching rows.
Eg: SQL> select * from t1 full outer join t2 on t1.a=t2.a and t1.b=t2.b;
A B C A B
------------------ -------------
X y z x y
P q r null null
Null null null s t

5. Natural Join: This join also returns matching rows only. When we are using
“Natural Join” then we are not allowed to used to use joining condition. In this case oracle
server only internally uses joining condition. But here resource tables must contain
common column name.
Syntax: select * from tablename1 natural join tablename2.
Note: Natural Join performance is very high compared to inner join.
Note: This join internally uses “USING” clause. That‟s why this join also returns common
columns one time only.
Eg: SQL> select * from t1 natural join t2;
Output:

A B C
-----------------
X y z
Note: When we are using “Natural” join also then we are not allowed to use alias name
on joining conditional column because natural join internally uses “USING” clause.
Eg: SQL> select ename,sal,deptno,dname,loc from emp e natural join dept d;

CROSS JOIN:
Eg: select ename,sal,dname, loc from emp cross join dept;
Joining more than 2-tables.(example 3 tables):
8i JOINS
Syntax: SQL> select col1,col2,….. from table1,table2,table3
Where table1.commoncolname=table2.commcolname
and table2. Commcolname=table3.commcolname;
9i JOIN:
Syntax: SQL> select col1,col2,… from table1 join table2
on table1.commcolname=table2.commcolname join table3
MUARLI SIR 2015
45
on table2.commcolname=table3.commcolname;

CONSTRAINTS:
Constraints are used to prevents or stops invalid data entry into our tables. Generally
constraints are created on table columns. Oracle server having following types of
constraints.
1. Not null
2. Unique
3. Primary key
4. Foreign key
5. Check

These are called “Constraints (or) Constraint types”.

Date: 3/4/15

All the above constraints are created in two ways:

1. Using column level

2. Using table level
1. Using Column level: In this methos we are defining constraints on individual
columns. That is whenever we are defining the column then only immediately we are
specifying constraint type.

Syntax: create table tablename(col1 datatype(size) constraint type, col2 datatype(size)

constraint type,…..);

2. Using Table level: In this method we are defining constraints on group of columns
i.e., in this method first we are defining all columns and last only we are specifying
constraint type along with group of columns.

Syntax: create table tablename(col1 datatype(size),col2 datatype(size),…. ,

constrainttype (col1,col2,….));

1. Not NULL: This constraint doesn‟t support table level. This constraint doesn‟t accepts
null values. But it will accepts duplicate values.

Column level: SQL> create table t1(sno number(10) not null, name varchar2(10));
Testing: SQL> insert into t1 values(null,‟abc‟);
ORA-1400: cannot insert null into SNO;
Table level: SQL> create table t1(sno number(10), name varchar2(10), not
null(sno,name));
2. UNIQUE: This constraint is defined on column level, table level. This constraint doesn‟t
accepts duplicate values. But it will accepts null values.
Note: Whenever we are creating unique constraints then oracle server internally
automatically creates b-tree index on those columns.
Column level: SQL> create table t2(sno number(10) unique, name varchar2(10));
MUARLI SIR 2015
46
Table level: SQL> create table t2(sno number(10), name varchar2(10),
unique(sno,name));
Table created
Eg: SQL> insert into t3(….)
SNO NAME
-----------------------
1 murali
1 abc
SQL> select * from t3;
Error: ORA-00001: Unique constraint violated.

3. Primary Key: Primary key “Uniquely identifying a record in a table”. There can be only
one primary key in a table and also primary key doesn‟t accepts duplicate null values.
Note: Whenever we are creating primary key then oracle server automatically create an
“b- tree” indexes on those columns.
Column level: SQL> create table t4(sno number(10) primary key, name varchar2(10));
Table created
Table level: SQL> create table t4(sno number(10), name varchar2(10), primary
key(sno,name));
This is also called as “Composite Primary Key” i.e., it is the combination of columns as a
“single primary key”.

4. Foreign Key: If we want to establishes relationship between tables then we are using
“Referential Integrity Constraints” (Foreign Key). One table foreign key must belongs to
another table “primary key”. Here these two columns must belong to same data types.
Always foreign key values based on “primary key” values only. Generally, primary key
doesn‟t accepts duplicate, null, values where as foreign key accepts duplicate, null values.

Date: 4/4/15

Column level (“References”):

Syntax: create table tablename(col1 datatype(size) references master tablename(primary
columnname),….);
Eg: SQL> create table h4(sno number(10) references t4);
Or
SQL> create table g4(x number((10) references t4(sno));

Table Level: (Foreign key references)

Syntax: create table tablename(col1 datatype(size),col2 datatype(size),…., foreign
key(col1,col2));
Eg: SQL> create table h5(sno number(10), name varchar2(10), foreign key(sno,name)
references t5);
Whenever we are establishing relationship between tables then oracle violates following
two rules:-
MUARLI SIR 2015
47
1. Deletion in “MASTER” table
2. Insertion in “CHILD” table

1. Deletion in “Master” table: When we try to delete master table record if child table
records are exist, then oracle server returns an error “ORA-2292” to overcome this
problem first we are deleting child table records in “child table” and then only we are
deleting appropriate “master table” records within master table. Otherwise using
“on delete cascade” clause.
ON DELETE CASCADE: When we are using this clause in child table then if we are
deleting a “master table” record within master table then oracle server automatically
deletes “master table” records in master table and also deletes corresponding records
in “child table”.
Syntax: create table tablename(col1 datatype(size) references mastertablename
(primary key colname) on delete cascade,…..);
Eg: SQL> create table mas(sno number(10) primary key);
SQL> insert into mas values(……..);
SQL> select * from mas;
SNO
----
1
2
3

SQL> create table child(sno number(10) references mas(sno) on delete cascade);

SQL> insert into child values(……..);
SQL> select * from child;
SNO
-----
1
1
2
2
3
3

Testing : (deletion in master table)

SQL> delete from mas where sno=1;
1 row deleted
SQL> select * from mas:
SQL> select * from child;
SNO SNO
---- ----
2 2
3 2
MUARLI SIR 2015
48
 3
3
Oracle also supports another clause along with foreign key “ON DELETE SET NULL”.
ON DELETE SET NULL:
Whenever we are using this clause the “child table” then if we deleting a master table
record then oracle server automatically deletes primary key value in master table and
also corresponding foreign key values are automatically set to null in child table.
Syntax:
create table tablename(col datatype(size)
reference master tablename(primary key colname) on delete set null,……);

2. INSERTION IN CHILD TABLE: In Oracle, when we are trying to insert other than
primary key values into Foreign key then oracle server returns an error “ORA-2291”.
Because in all database systems always foreign key values are based on primary key
values only.
Eg: SQL> insert into child values(5);
ORA-2291: Intergrity constraint violated-parent key not found.
Solution: SQL> insert into mas values(5);
SQL> select * from mas;
SQL> insert into child values(5);
SQL> select * from mas;
Example based on dept(master table), emp(child table):
1) Deletion in Master table:
SQL> delete from dept where deptno=10;
ORA-2292: integrity constraint violated- child record foun.
2) Insertion in child table:
SQL> insert into emp(empno,deptno) values(5,90);
ORA-2291: Integrity constraint violated – parent key not found

Date: 6/4/15

In oracle, we are not allowed to define logical condtions using “SYSDATE” within check
constraint.
COLUMN LEVEL:
Syntax: create table tablename(col1 datatype(size) check(logical condition),col2
datatype(size),…..);
Eg: create table t6(sal number(10) check (sal>5000));
SQL> insert into t6 values(3000);
Error: check constraint(scott.syst_c005513) violated
SQL> insert into t6 values(9000);
SQL> select * from t6;
SAL
-----
9000
MUARLI SIR 2015
49
Eg: create table t7(name varchar2(20) check(name=upper(name)));
SQL> insert into t7 values(„murali‟);
Error: check constraint(scott.syst_c00514) violated
SQL> insert into t7 values („MURALI‟);
SQL> select * from t7;
NAME
------
MURALI
Table level:
Eg: SQL> create table t8(name varchar2(20), sal number(10), check(name=upper(name)
and sal>5000));
Assign User defined names to Constraints (OR) User defined Constraint Names:
In all database systems whenever we are creating constraints then database servers
internally automatically generates an unique identification number for indentifying a
constraint uniquely.
In Oracle also whenever we are creating constraints then oracle server automatically
generates an unique identification number in the format of “SYS_Cn”. This is called
“Predefined Constraint Name”. In place of this one we can also create our own name using
“Constraint Keyword”. This is called “Used defined Constraint Name”.
Syntax: Constraint user defined name constraint_type;
Here, constraint-> keyword
Constraint_type-> Not null, unique, primary key, foreign key, check
User defined name->constraint name
Eg 1: (Predefined Constraint Name)
SQL> create table t10(sno number(10) primary key;
Testing:
SQL> insert into t10 values(1);
SQL> insert into t10 values(1);
Error: unique constraint (SCOTT.SYS_C005516) violated
Eg 2: User defined Constraint Name
SQL> create table t11 (sno number(10) constraint p_abc primary key);
Testing:
SQL> insert into t11 values(1);
SQL> insert into t11 values(1);
Error: unique constraints (SCOTT.P_ABC) violated
Data Dictionaries:
Whenever we are installing oracle server then oracle server automatically creates some
read only tables. These read only tables are also called as “Data Dictionaries”.

MUARLI SIR 2015

50
Note 1: In Oracle, all constraints information stored under “user_constraint” data
dictionary.
Eg: SQL> desc user_constraints;
SQL> select CONSTRAINT_NAME, CONSTRAINT_TYPE from USER_CONSTRAINTS where
table_name=‟EMP‟;
Output:
CONSTRAINT_NAME CONSTRAINT_TYPE
-------------------------- --------------------------------------
PK_EMP P
FK_DEPTNO R
 Table_name=‟EMP‟

Here, EMP is capital letter

NOTE 2: In Oracle, if you want to view column names along with constraint names then
we are using “user_cons_columns”.
Eg: SQL> desc user_cons_columns;
SQL> select constraint_name,column_name from user_cons_columns
where table_name=‟EMP‟;
Output:
CONSTRAINT_NAME COLUMN_NAME
------------------------ -------------------
FK_DEPTNO DEPTNO
PK_EMPNO EMPNO
NOTE 3: In Oracle, if you want to view logical conditions of the check constraint then we
are using “search_condition” from “user_constraints” data dictionary.
Eg: SQL> desc user_constraints;
SQL> select search_condition from user_constraints
where table_name=‟T7‟;
SEARCH_CONSTRAINTS
---------------------------
name=upper(name)

Date: 7/4/15

In Oracle, all columns information stored under “user_tab_columns” data dictionary.

MUARLI SIR 2015
51
SQL> desc user_tab_columns;
SQL> select column_name from user_tab_columns
where table_name=‟EMP‟;
Q) Write a query to display number of columns from emp table?
Ans: SQL> select count(*) from user_tab_columns where table-name=‟EMP‟;

DEFAULT: In Oracle, we can also provide “default” values for a column using “DEFAULT”
constraint.
Syntax: columnname datatype(size) default actual value;
Eg: SQL> create table b1(name varchar(10), sal number(10) default 5000);
SQL> insert into b1(name) values („abc‟);
SQL> select * from b1;
NOTE: In Oracle, if you want to view default values for a column then we are using
“data_default” property (predefined column) from “user_tab_columns” data dictionary.
Eg: SQL> desc user_tab_columns;
SQL> select column_name,data_default from user_tab_columns where table_name=‟B1‟;
COLUMN_NAME DATA_DEFAULT
------------------------------------------------
SAL 5000
Adding (or) Dropping constraints on existing table:
In Oracle, we can add or drop constraints on existing table using “ALTER” command.
NOTE 1: If we want to add constraints on existing table or existing column then we are
using “table level syntax method”.
Eg: SQL> alter table a1
add primary key(sno);

NOTE 2: If we want to add a new column along with constraint then we are using
“Column level syntax method”.
Eg: SQL> alter table a1 add name varchar2(10) unique;

MUARLI SIR 2015

52
Q) Add foreign key in existing a2 references a1?
Ans: SQL> alter table a2 add foreign key(sno) references a1(sno);

NOTE 3: If we want to add “NOT NULL” constraints into existing table or existing column
then we are using “alter with modify”.
Syntax: alter table tablename modify columnname not null;
Eg: SQL> create table a3(sno number(10));
SQL> alter table a3 modify sno not null;

Q) Write a query to add new column with not null into a3?
Ans: SQL> alter table a3 add name varchar2(10) not null;

NOTE: In Oracle, whenever we are copying a table from another table constraints are
never copied. But in all database systems “not null” constraints only copied.
Eg: SQL> create table a4 as select * from a3;
“not null is copied”
SQL> desc a4;
Eg: SQL> create table dept1 as select * from dept;
SQL> select * from dept1;
SQL> alter table dept1 add primary key(deptno);
SQL> create table emp1 as select * from emp;
SQL> select * from emp1;
SQL> alter table emp1 add primary key(empno);
SQL> alter table emp1 add foreign key deptno references emp1(deptno) on delete
cascade;
Dropping Constraints: By dropping “Constraint name” two methods for dropping
constraints.
Method 1: Alter table tablename drop constraint constraint name;
OR
Method 2:
A. Alter table tablename drop primary key;
B. Alter table tablename drop unique(col1,col2,….);

MUARLI SIR 2015

53
Date: 8/4/15

SQL> create table test(sno number(10) primary key);

SQL> alter table test drop primary key;
NOTE: In Oracle, if we want to drop primary key along with referenced foreign keys then
we are using “Cascade” clause along with “Alter drop”.
Syntax: alter table tablename drop primary key cascade;
SQL> alter table a1 drop primary key;
Error: this unique/primary key is referenced by some foreign key.
Solution: SQL> alter table a1 drop primary key cascade;
Eg: Not Null constraints for dropping.
SQL> desc a3;
SQL> desc user_cons_columns;
SQL> select column_name,constraint_name from user_cons_columns where
table_name=‟A3”;
COLUMN_NAME CONSTRAINT_NAME
------------------------------------------------------
NAME SYS_C005526
SNO SYS_C005525
SQL> alter table a3 drop constraint SYS_C005526;
SQL> alter table a3 drop constraint SYS_C005525;
SQL> desc a3;

SUB QUERY (or) NESTED QUERY:

Query within another query is called “Sub Query” or “Nested Query”. Sub query‟s are used
to retrieve data from “single” or “multiple” tables based on “More than one step process”.
All relational data base having two types of sub query‟s:
1. Non- Correlated sub query‟s
2. Correlated sub query‟s
In Non-correlated sub query‟s “child” query is executed first then only “parent” query is
executed, whereas in correlated sub query‟s “parent” query is executed first then only
“child” query is executed.
1. Non-Correlated sub query‟s: These sub query‟s mainly consist of two parts.
I. Child query
II. Parent query
1. Child Query: A query which provides values to the another query is called “Child
Query or Inner Query”.
2. Parent Query: A query which receives values from another query is called “Parent
Query or Outer Query”.
In non-correlated sub query maximum limit is up to “255” query‟s.

NON-CORRELATED SUB QUERY:

 Single row sub query‟s
MUARLI SIR 2015
54
 Multiple row sub query‟s
 Multiple column sub query‟s
 Inline view or sub query‟s are using in “from” clause.

Q) Write a query to display the employees who are getting more salary than the avg(sal)
from emp?
Ans: SQL> select ename, sal from emp where sal>(select avg(sal) from emp);
This is a “Single row sub query” because here child query returns single value. In Single
row sub query‟s we are using =,<,>,<=,>=, between operators.
Execution:
Step 1: SQL> select avg(sal) from emp;
2830.35714
Step 2: SQL> select * from emp where sal>2830.35714

Q) Write a query to display the employees who are working in sales department from
emp, dept tables.
Ans: SQL> select ename from emp where deptno=(select deptno from dept where
dname=‟SALES‟);

Date: 9/4/15

Q) Write a query to display the employees who are working in “SMITH” department
number from emp table?
Ans: SQL> select ename from emp where deptno=(select deptno from emp where
ename=‟SMITH‟);

Q) Write a query to display senior most employee details from emp table?
Ans: SQL> select * from emp where hiredate=(select min(hiredate) from emp);

Q) Write a query to display highest sal details emp table?

Ans: SQL> select * from emp where sal=(select max (sal) from emp);

Q) Write a query to display highest paid emp department, dname from emp, dept tables?
Ans: SQL> select dname from dept where deptno=(select deptno from emp where
sal=(select sal(max) from emp));

Q) Write a query to display second max(sal) from emp table?

Ans: SQL> select max(sal) from emp where sal<(select max(sal) from emp);

Q) Write a query to display second highest sal employee details from emp table?
Ans: SQL> select * from emp where sal=(select max(sal) from emp where sal<(select
max(sal) from emp));

MUARLI SIR 2015

55
Q) Write a query to display the employees who are working under „BLAKE‟ from emp table
using „empno, mgr‟ columns ?
Ans: SQL> select * from emp where mgr= (select empno from emp where
ename=‟BLAKE‟);

Q) Write a query to display lowest average salary job from emp tables?
Ans: SQL> select job, avg(sal) from emp
group by job
having avg(sal)=(select min(avg(sal)) from emp);
Error: nested group function without GROUP BY

NOTE: In all database systems whenever child query contains nested group functions
then we must use “Group BY” clause within child query.
Ans: SQL> select job, avg(sal) from emp group by job
having avg(sal)= (select min(avg(sal)) from emp group by job);
JOB AVG(SAL)
--------------------------
SALESMAN 1050

Q) Write a query to display lowest average salary job from emp table where deptno less
than 30?
Ans: SQL> select deptno, min(sal) from emp
group by deptno
having min(sal)>(select min(sal) from emp where deptno= 30);
DEPTNO MIN(SAL)
----------------------------------
10 4400
20 1700

Date: 10/4/15

Q) Write a query to display job, avg(sal) of the employees whose job avg(sal) more than
the clerk‟s job avg(sal) from emp table?
Ans: SQL> select avg(sal), job from emp group by job
having avg(sal) >(select avg(sal) from emp where job=‟CLERK‟);

NOTE: In all databases, whenever we are using sub query‟s always database servers
returns parent query table columns. i.e., we are unable to use child query table columns
into parent query to overcome this problem we must use joins within “Parent Query”.

Q) Write a query to display the employees who are working sales department from emp,
dept tables?
Ans: SQL> select ename, dname from emp e, dept d where e.deptno=d.deptno and
d.deptno=(select deptno from dept where dname=‟SALES‟);
MUARLI SIR 2015
56
ENAME DNAME
--------------------------------
ALLEN SALES
WARD SALES
MARTIN SALES…….

Q) Write a query to display the employees who are getting more salary than the „BLAKE‟
salary from emp table?
Ans: SQL> select * from emp where sal> (select sal from emp where ename=‟BLAKE‟);

MULTIPLE ROW SUB QUERY‟S:

Whenever child query returns multiple values those types of sub query‟s are called
„Multiple row Sub Query‟s‟.
Q) Write a query to display the employee details who are getting max(sal) in each
department from emp table?
Ans: SQL> select * from emp where sal= (select max(sal) from emp group by deptno);
Error: single row sub query returns more than one row.

This is an multiple row sub query because here query returns multiple values. In multiple
row sub query‟s we are using „in‟, „all‟, „any‟ operators.

NOTE: In all database systems we can also use „in‟ operator in single row sub query‟s.

Solution: SQL> select * from emp where sal in(select max(sal) from emp group by
deptno); OR
SQL> select ename, sal, deptno from emp where sal in(select max(sal) from emp group
by deptno);

Q) Write a query to display the employees who are working in sales or research
department from emp, dept?
Ans: SQL> select ename, deptno from emp where deptno in(select deptno from dept
where dname=‟SALES‟ or dname=‟RESEARCH‟);

Q) Write a query to display the employees who are working as “supervisors” (managers)
from emp table using empno, mgr?
Ans: SQL> select * from emp where empno in(select mgr from emp);

Q) Write a query to display the employees who are not working as “supervisors”
(managers) from emp using empno, mgr?
Ans: SQL> select * from emp where empno not in(select nvl(mgr,0) from emp);

MUARLI SIR 2015

57
Date: 11/4/15

TOP N Analysis:
1. Inline view
2. Rownum

1. Inline view: Oracle 7.2 introduced “Inline views”. In Inline views we are using “sub
query‟s” in place of tablename with in parent query.
Syntax: select * from (select statement);
Generally, in all database systems we are not allowed to use “Order By” clause in child
query‟s. To overcome this problem oracle 7.2 introduced sub query‟s in from “from”
clause. These type of query‟s are also called as “ INLINE QUERY‟s”.
Generally, in oracle we are not allowed to use “Alias name” in “where” condition. To
overcome this problem also we are using “Inline views”. Whenever we are using “Inline
views” this alias name automatically behaves like column name, then only this alias name
we can use in “where” condition.
Eg: SQL> select ename, sal, sal*12 annsal from emp where annsal>30000;
Error: ANNSAL: Invalid Idenfier.
Solution: SQL> select * from(select ename,sal,sal*12 annsal from emp) where
annsal>30000;
2. ROWNUM: (Magic Column) Rownum is an Psuedo column(generalised column) it
behaves like a table column. Whenever we are installing oracle server then
automatically some psuedo columns are created in database. These “Pseudo Columns”
belongs to all tables within database. Rownum psuedo column is used to filter the data
within a table.
Rownum is a psuedo column which automatically assigns numbers to each row in a table
at the time of selection.

SQL> select rownum, ename from emp;

ROWNUM ENAME
--------------------------------
1 SMITH
2 ALLEN
3 WARD

Generally, rownum having temporarily values.

Eg: SQL> select rownum, ename from emp where deptno=10;

ROWNUM ENAME
-------------------------------
1 CLARK
2 KING
MUARLI SIR 2015
58
3 XXX

Q) Write a query to display first row from emp table using rownum?
Ans: SQL> select * from emp where rownum=1;

Q) Write a query to display second row from emp table using rownum?
Ans: SQL> select * from emp where rownum=2;
No rows selected
Generally, rownum doesn‟t work with more than “1” +ve integer, i.e., it works with <, <=
operators.

Q) Write a query to display first five rows from emp table using rownum?
Ans: SQL> select * from emp where rownum<=5;

Q) Write a query to display first five highest salary employees from emp using rownum?
Ans: SQL> select * from( select * from emp order by sal desc) where rownum<=5;

Q) Write a query to display 5th highest salary employee from emp table using rownum?
Ans: SQL> select * from(select * from emp order by sal desc) where rownum=5
minus
select * from(select * from emp order by sal desc) where rownum<=4;

Date: 16/4/15

Q) Write a query to display rows between 1 to 5 from emp table using “Rownum”?
Ans: SQL> select * from emp where rownum between 1 and 5;

Q) Write a query to display rows between to 5 to 9 from table?

Ans: SQL> select * from emp where rownum between 5 and 9;
No rows selected
Solution: SQL> select * from emp where rownum <=9
minus
select * from emp where rownum<=5;

Q) Write a query to display second record from emp using rownum?

Ans: SQL> select * from emp where rownum<=2
minus
select * from emp where rownum<=1;

Q) Write a query to display last two records from emp table using rownum?
Ans: SQL> select * from emp where rownum<=14
minus
select * from emp where rownum<=12;
Solution: SQL> select * from emp minus select * from emp where rownum<=
MUARLI SIR 2015
59
(select count(*)-2 from emp);

 SQL> select * from emp where rownum>1;

No rows selected
 SQL> select * from emp where rownum>=1;
Rows selected

NOTE: Whenever we are using alias name for rownum in “INLINE VIEWS” then that alias
name works with all SQL operators because that alias name behaves like a table column.

Q) Write a query to display second record from emp table using “rownum” alias name?
Ans: SQL> select * from(select rownum r, ename, job, sal from emp) where r=2;
R ENAME JOB SAL
------------------------------------------
2 ALLEN SALESMAN 1400

NOTE: If we want to display all the columns from the table using rownum alias name then
we must use table.star(*) i.e., {e.*} along with rownum alias name after select.
Otherwise create an alias name for the table and use aliasname.* after select along with
rownum alias name.
SQL> select * from( select rownum r, emp.* from emp) where r=2;
OR
SQL> select * from(select rownum r, e.* from emp e) where r=2;

Q) Write a query to display second, third, fifth, seventh, eighth rows from emp table using
rownum alias name?
Ans: SQL> select * from (select rownum r, e.* from emp e) where r in(2,3,5,7,8);

Q) Write a query to display records from 5 to 9 from emp table using rownum alias name?
Ans: SQL> select * from(select rownum r, emp.* from emp) where r between 5 and 9;

Q) Write a query to display first and last records from emp table using rownum alias
name?
Ans: SQL> select * from(select rownum r, emp.* from emp) where r=1 or r=(select
count(*) from emp);

Q) Write a query to display even number of records from emp table using rownum alias
name?
Ans: SQL> select * from(select rownum r, emp.* from emp) where r in(select mod(r,2)=0
from emp);
Solution: SQL> select * from(select rownum r, emp.* from emp) where mod(r,2)=0;

Q) Write a query to display skip first 5 rows from emp table using rownum alias name?
Ans: SQL> select * from (select rownum r, emp.* from emp) where r>5;
MUARLI SIR 2015
60
ANALYTICAL FUNCTIONS are used in INLINE VIEWS.
ANALYTICAL FUNCTIONS:
Oracle 8i introduced “Analytical Functions”. These functions are similar to “Group
functions” but group functions reduces number of rows within the group. Where as
“Analytical Functios” doesn‟t reduces number of rows with in the group. That is when we
are using analytical functions database servers executes for each and every row with in
the group.
Using GROUP function.
Eg: SQL> select deptno, avg(sal) from emp group by deptno;
DEPTNO AVG(SAL)
---------------------------
10 5883.3333
20 2935
30 1666.66667

Using Analytical Functions

SQL> select deptno, avg(sal) over(partition by deptno) from emp;

Oracle having following Analytical functions.

1. Row_number()
2. Rank()
3. Dense_rank()

These three analytical functions automatically assigns “ranks” to each row in a table either
group wise or rows wise in a table.

Date : 17/4/15

Analytical Function:
Syntax:
Analytical function name over(partition by columnname order by columnname[asc/desc]);
Row_number() analytical function automatically assigns different rank numbers when
values are same. Where as rank() and dense_rank() analytical functions automatically
assigns same rank numbers when values are same. And also rank() skips next
consecutive rank numbers where as dense_rank() doesn‟t skips next consecutive rank
numbers.

Eg: row_number():
DEPTNO ENAME SAL R
-----------------------------------------
20 SCOTT 3800 1
20 FORD 3800 2
20 JONES 3475 3
MUARLI SIR 2015
61
Eg: rank():
DEPTNO ENAME SAL R
-----------------------------------------
20 SCOTT 3800 1
20 FORD 3800 1
20 JONES 3475 3

Eg: dense_rank():
DEPTNO ENAME SAL R
-----------------------------------------
20 SCOTT 3800 1
20 FORD 3800 1
20 JONES 3475 2

Q) Write a query to display highest salaries to lowest salaries in each department from
emp table using row_number() analytical function?
Ans: SQL> select deptno, ename,sal, row_number over(partition by deptno order by sal
desc) r from emp;

Q) Write a query to display 2nd highest salary employee in each department from emp
table using analytical function?
Ans: SQL> select * from(select deptno, ename, sal, dense_rank() over(partition by
deptno order by sal desc)r from emp) where r=2;

Q) Write a query to display 5th highest salary employee from emp table using analytical
function?
Ans: SQL> select * from(select deptno, ename, sal, dense_rank() over(order by sal
desc)r from emp) where r=5;

NOTE: In analytical functions “Partition By” clause is an optional clause.

Q) Write a query to display nth highest salary employee from emp table using analytical
function?
Ans: SQL> select * from(select deptno, ename, sal, dense_rank() over(order by sal
desc)r from emp) where r=&n;

Output: Enter value for n: 1

DEPTNO ENAME SAL R
--------------------------------------
10 KING 7900 1

MUARLI SIR 2015

62
ROWID:
Rowid is an pseudo column, it behaves like a table column whenever we are inserting data
into table then oracle server automatically generates an unique identification number for
identifying a record uniquely. This is called “ROWID”. Generally, using rowid‟s we can
retrieve data very quickly from database.
Generally, rownum having temporary values where as rowed having fixed values.
Eg: SQL> select rownum, rowid, ename from emp;

Eg: SQL> select rownum, rowid, ename from emp where deptno=10;
In Oracle, by default rowid‟s are ascending order.

Q) Write a query to display second record from emp table using analytical function, rowid?
Ans: select * from(select deptno, ename, sal, row_number() over(order by rowid)r from
emp) where r=2;

Q) Write a query to display each department second row from emp table using analytical
function, rowid?
Ans: SQL> select * from(select deptno, ename, sal, row_number() over (partition by
deptno order by rowid) r from emp where r=2);

Q) Write a query to display last two records from emp table using analytical function,
rowid?
Ans: SQL> select * from(select deptno, ename, sal, row_number() over (order by rowid
desc) r from emp)where r<=2;

Date: 18/4/15

We can also use max(), min() functions in “rowid‟s”.

Eg: SQL> select max(rowid) from emp;
SQL> select max(rowid) from emp;

Q) Write a query to display first row from emp table using rowid?
Ans: SQL> select * from emp where rowid=(select min(rowid) from emp);

Q) Write a query to display last record from emp table using rowid?
Ans: SQL> select * from emp where rowid=(select max(rowid) from emp);

Q) Write a query to display duplicate data from the following table?

SQL> create table test(sno number(10));
SQL> insert into test values(&sno);
Enter values for sno: 10
Enter values for sno: 10
Enter values for sno: 10
Enter values for sno: 20
MUARLI SIR 2015
63
Enter values for sno: 20
Enter values for sno: 30
Enter values for sno: 40
SQL> select * from test;
SNO
-----
10
10
10
20
20
30
40

SQL> select sno, count(*) from test group by sno having count(*)>1;
SNO COUNT(*)
----------------------------
10 3
20 2

Generally, rowid‟s are used in subquery‟s for deleting duplicate rows in a table.
Q) Write a query to delete duplicate rows except one row in each group from a table using
rowid?
Ans: SQL> delete from test where rowid not in(select min(rowid) from test group by sno);
SQL> select * from test;
SNO
-----
10
20
30
40
Multiple Column Sub Query‟s:
In all database we can also compare multiple columns of the child query with the multiple
columns of the parent query these type of query‟s are also called as “Multiple Column Sub
Query‟s”.
In all database systems when we are using multiple column sub query‟s then we must
specify parent query where conditional column‟s within paranthesis “()”.
Syntax: select * (or) col1, col2,…. from tablename where(col1, col2,…….) in (select col1,
col2,…… from tablename where condition).

Q) Write a query to display the employees whose job, mgr match with the job, mgr of the
employee „SCOTT‟ from emp table?
Ans: SQL> select * from emp where(job,mgr) in(select job, mgr from emp where
ename= „SCOTT‟);
MUARLI SIR 2015
64
Q) Write a query to display highest salary employees in each department from emp table
using multiple row sub querys?
Ans: SQL> select deptno, sal, ename from emp where sal in (select max(sal) from emp
group by deptno);

DPETNO SAL ENAME

-----------------------------------------
10 7900 KING
20 3800 FORD
20 3450 SCOTT
30 3450 BLAKE

Solution:
Q) Write a query to display the employees who are getting maximum salary in each
department from emp table using multiple column sub query?
Ans: SQL> select deptno, sal, ename from emp where(deptno, sal) in (select deptno,
max(sal) from emp group by deptno);
DPETNO SAL ENAME
-----------------------------------------
10 7900 KING
20 3800 FORD
30 3450 BLAKE

Q) Write a query to display senior most employees in each job from emp table using
multiple column sub query?
Ans: SQL> select hiredate, ename, job from emp where (hiredate, job) in (select
min(hiredate), job from emp group by job);

Q) Write a query to display ename, dname, sal of the employees whose sal, comm match
with the sal, comm. Of the employees working in the location „DALLAS‟?
Ans: SQL> select ename, dname, sal from emp e, dept d where e.deptno=d.deptno
and (sal,nvl(comm,0)) in (select sal, nvl(comm,0) from emp where deptno=(select
deptno from dept where loc=‟DALLAS‟));

Date: 20/4/15

Q) Write a query to display the employee who are getting more salary than the highest
paid employee in 20th department from emp table?
Ans: SQL> select * from emp where sal> (select max(sal) from emp where deptno=20);

Q) Write a query to display the employees who are getting more than the lowest paid emp
in 10th dept?
Ans: SQL> select * from emp where sal> (select min(sal) from emp where deptno=10);
MUARLI SIR 2015
65
Whenever resouce table having large amount of data and also child query contains max()
or min() functions and also when we are comparing then those type of sybquery‟s
degrades performance of the application. To overcome this problem to improve
performance of these type of query‟s ANSI/ISO SQL. These are “all”,”any”. These sub
query‟s special operators are used along with relational operators in parent query “where”
condition.

SQL> select * from emp where sal> all(select sal from emp where deptno=20);
SQL> select * from emp where sal> any(select sal from emp where deptno=10);

Sub Query Special Operators are used in Multiple Row Sub Query:
When child query returns multiple values then those type sub query are called “Multiple
Row Sub Query‟s”. In all database systems “in”, ”all”, “any” operators used in “Multiple
Row Sub Query‟s”.
In-> It returns same values in the list(child query)
All-> It satisfies all values in the list
Any-> It satisfies any value in the list
IN-> =ANY
Q) Write a query to display the employees who are getting more than the all salaries of
the clerks from emp table using sub query special operator?
Ans: SQL> select * from emp where sal> all(select sal from emp where job=‟CLERK‟);
 To display the result in order(asc/desc)

SQL> select * from emp where sal> all(select sal from emp where job=‟CLERK‟) order by
sal desc;

Note: whenever we are using sub query special operator “OR” internally database servers
uses logical operator “AND”
Eg: SQL> select * from emp where deptno>all(10,20);
Output: 30
Note: Whenever we are using sub query special operator “ANY” internally database
servers logically operators “OR”
Eg: SQL> select * from emp where deptno>any(10,20);
Output: 20
30
NOTE: “=any” is also same as “in” operator but “<>any” is not same as “not in” operator.
Eg: SQL> select * from emp where deptno not in(10,20)
Output: 30

SQL> select * from emp where deptno <> any(10,20);

Output: 10
20
30
MUARLI SIR 2015
66
NOTE: “<>all” is also same as “not in” operator.
Eg: SQL> select * from emp where deptno not in(10,20);
SQL> select * from emp where deptno <>all(10,20);

CORRELATED SUB QUERY‟S:

In Non-Correlated sub query‟s „child query‟ is executed first then only „parent query‟ is
executed. Where as in correlated sub query‟s „parent query‟ is executed first then only
„child query‟ is executed. Whenever we are submitting correlated sub query then database
servers get a candidate row from the „parent query‟ table and then control passed into
„child query‟ where condition and then based on evaluation value it compare with „parent
query‟ where condition. Generally, in non-correlated sub query „child query‟ is executed
only once per „parent query‟ table where as in correlated sub query.
„Child query‟ is executed for each row for „parent query‟ table. In all database systems in
correlated sub query we must create alias name for „parent query‟ table within „parent
query‟ and use that alias name in child query “where condition”.

Syntax: select * from tablename aliasname when columnname = (select * from

tablename where column name =(any operator) aliasname.columnname);

Generally, correlated sub query are used in denormalization process in denormalization

process in this process we are using correlated updates. Generally, if you want to modify
one table column values based on another table and also those tables are related then
only we are using “ Correlated update”.

Syntax: update tablename1 aliasname1 set columnname = (select columnname from

tablename2 aliasname2 where aliasname1.common col name);
SQL> alter table emp add dname varchar2(10);
SQL> update emp e set dname=(select dname from dept d where e.deptno=d.deptno);
SQL> select * from emp;

Date: 21/4/15

Q) Write a query to display the employees who are getting more salary than the average
salaries of their job from emp using correleated sub query?
Ans: SQL> select * from emp e where sal> (select avg(sal) from emp where job=e.job);

MUARLI SIR 2015

67
Non-Correlated Sub Query:

Correlated Sub Query:

Q) Write a query to display first highest salary employee from emp table using correlated
sub query??
Ans: SQL> select * from emp e1 where 1 =(select count(*) from emp e2 where
e1.sal>=e2.sal);
Output: king 7900

Q) Write a query to display second highest salary employee from following table using
correlated sub query?
Ans: SQL> create table test(ename varchar(20), sal number(10));
SQL> insert into test values(„&e‟,&s);
Enter the values for e: abc
Enter the values for s: 100
SQL> /
Enter the values for e: xyz
Enter the values for s: 150
SQL> /
Enter the values for e: zzz
Enter the values for s: 200
SQL> /
Enter the values for e: bbb
Enter the values for s: 300

SQL> select * from test;

ENAME SAL
MUARLI SIR 2015
68
------------------------
abc 100
xyz 150
zzz 200
bbb 300

SQL> select * from test e1 where 2=(select count(*) from test e2 where e2.sal>=e1.sal);

Execution Process:
Phase 1:
Step 1: get a “Candidate Row” [where cursor pointer points a record is called “Candidate
Row”] (abc 100)
Step 2: select count(*) from test e2 where e2.sal>=100;
4
Step 3: select * from e1. Where 2=4;
False

Phase 2:
Step 1: get a “Candidate Row” [where cursor pointer points a record is called “Candidate
Row”] (xyz 150)
Step 2: select count(*) from test e2 where e2.sal>=150;
3
Step 3: select * from e1. Where 2=3;
False

Phase 3:
Step 1: get a “Candidate Row” [where cursor pointer points a record is called “Candidate
Row”] (zzz 200)
Step 2: select count(*) from test e2 where e2.sal>=200;
2
Step 3: select * from e1. Where 2=2;
True

Output: zzz 200

NOTE: Whenever resource table having duplicate data based on where conditional column
then the above query doesn‟t return any rows. To overcome this problem we must use
“DISTINCT” clause within “Count(*)” function.
Eg: SQL> insert into test values(„murali‟,200);
SQL> select * from test;
ENAME SAL
------------------------
abc 100
xyz 150
MUARLI SIR 2015
69
zzz 200
bbb 300
murali 200

SQL> select * from test e1 where 2=(select count(*) from test e2 where e2.sal>=e1.sal);
No Rows Selected
Solution:
SQL> select * from test e1 where 2=(select count(distinct(sal)) from test e2 where
e2.sal=e1.sal);
Output:
Zzz 200
Murali 200

NOTE: We can also write above query using “n-1” method. In this case we are not allowed
to use “=” operator in “Where” condition of “Child Query”.

Eg: SQL> select * from test e1 where (2-1)=(select count(distinct(sal)) from test e2
where e2.sal>e1.sal);

Q) Write a query to display nth highest salary employee from emp table using “Correlated
Subquery”.?
Ans: SQL> select * from emp e1 where &n = (select count(distinct(sal)) from emp e2
where e2.sal>=e1.sal);

Date : 22/4/15

Exists Operator: Exists operator used in “Correlated sub query‟s”. Exists Operator
performance is very high. Compare to “in” operator. Exists operator always returns
“Boolean Values” either “true” or “false”. Exists Operator is used in “where” condition of
the “Parent Query” only and also when we are using Exists Operator then we are not
allowed to use “Column Name” along with “Exists Operator” in “Where” Condition.
Syntax:
Select * from tablename aliasname where exists(select * from tablename where
columnname=aliasname. Columnname);
In all database systems if we want to test whether one table values are availabe or not
availabe in another table columns then only we are using “Exists Operator”.

NOTE: Exists Operator is used to test whether a given set is empty (or) non-empty Exists
on non-empty set returns “true” whereas exists operator on empty set returns “false”.
Eg 1: exists {1,2,3}= true
Eg 2: exists { }= false

Q) Write a query to display those departments from dept table having employees in emp
table using correlated sub query exists operator?
MUARLI SIR 2015
70
Ans: SQL> select * from dept d where exists(select * from emp where deptno=d.deptno);
Output:
DEPTNO DNAME LOC
-------------------------------------------------
10 ACCOUNTING NEW YORK
20 RESEARCH DALLAS
30 SALES CHICAGO

Q) Write a query to display the employees who are getting same salary as scott salary
from emp table using correlated sub query exists operator from emp?
Ans: SQL> select * from emp e1 where exists(select * from emp e2 where
ename=‟SCOTT‟ and e2.sal=e1.sal);
Output:
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPNTO
-------------------------------------------------------------------------------------------------
7902 FORD ANALYST 7566 03-DEC-81 3000 20
7788 SCOTT ANALYST 7566 19-APR-87 3000 20

 For not displaying „SCOTT‟ in the output table:

Solution: SQL> select * from emp e1 where exists(select * from emp e2 where
ename=‟SCOTT‟ and e2.sal=e1.sal);

Output:
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPNTO
-------------------------------------------------------------------------------------------------
7902 FORD ANALYST 7566 03-DEC-81 3000 20

Q) Write a query to display those departments from dept table doesn‟t have employees in
emp table using correlated sub query?
Ans: SQL> select * from dept d where not exists(select * from emp where
deptno=d.deptno);
Output: 40 operations boston

Q) Write a query to display those department doesn‟t have employees in emp table using
non-correlated sub query‟s?
Ans: SQL> select * from dept where deptno not in(select deptno from emp);
Output: 40 operations boston

Generally, “not in” operator doesn‟t work with null values to overcome this problem we
are using “not exists” operator along with “correlated sub query”.
SQL> select * from emp;
SQL> select * from dept where deptno not in(select deptno from emp);
MUARLI SIR 2015
71
No Rows Selected
SQL> select * from dept d where not exists(select * from emp where deptno=d.deptno);
Output: 40 Operations BOSTON

Date: 23/4/15

NOTE: In Oracle, whenever we are testing one table column values are available (or) not
available in another table using not correlated and correlated query‟s. These type of
query‟s are also called as “Anti Joins”.
Eg: SQL> select * from dept where deptno not in(select deptno from emp);

VIEWS
View is an database object which is used to provide authority level of security. Generally,
views are created by “Database Administrator”.
View doesn‟t store data that‟s why view is also called as “Virtual table”. View is also called
as “Window of a table”.
Generally views are created from “Base tables”. Based on the type of base tables views
are categorized into two views.
1. Simple View
2. Complex view (or) Join view

“Simple view” is a view which is created from only “One Base” tables where as
“Complex view” is a view which is created from “Multiple base” tables.

Simple View:
Syntax: create or replace view viewname as
select * from tablename where condition;

DML Operations performed on Simple view:

1. When a Simple view contains “Group functions”, “Group by”, ”Rownum”, “Distinct”,
“Set Operators”, “Joins” then we cannot perform DML operation through “Simple view” to
“Base table”.
2. We must include “Base table” not null column into the view then only we can
perform “Insertion operation” through “View2” base table. Otherwise oracle server
returns an error.

Eg 1: SQL> create or replace view v1 as select * from emp where deptno=10;

SQL> select * from v1;
SQL> insert into v1(empno,ename,deptno) values(1,‟murali‟,30);
1 row created
SQL> select * from emp;

MUARLI SIR 2015

72
Eg 2: SQL> create or replace view v2 as select ename, sal, deptno from emp where
deptno=10;
SQL> select * from v2;
SQL> insert into v2(ename, sal, deptno) values(„abc‟,1000,30);
Error: Cannot insert null into empno.
In Oracle, whenever we are using “Group functions” or “rownum” in a view then we must
create alias name for those expressions otherwise oracle server returns an error.
Eg: SQL> create or replace view v3 as select depnto, max(sal) a from emp group by
deptno;
View created
SQL> select * from v3;
SQL> create or replace view v4 as select rownum, ename from emp;
Error: must name this expression with column alias.
SQL> create or replace view v4 as select rownum rno, ename from emp;
SQL> select * from v4;
In every database systems whenever we are creating a view then automatically view
definitions (SELECT STATEMENTS…….) are automatically permanently stored in database.
In Oracle, if we want to view these definitions then we are using “user_views” data
dictionary.
Eg: SQL> desc user_views;
SQL> select text from user_views where view_name=‟V3‟;
Output:
TEXT
-------------------------------------------------------------------------
select deptno,max(sal) a from emp group by deptno;

Date: 24/4/15

Complex View: Complex view is a view which is created from multiple base tables.
Eg: SQL> create or replace view v5
as
select ename, sal,dname, loc
from emp,dept
where emp.deptno=dept.deptno;
SQL> select * from v5;

DML Operations on Complex Views:

Eg: SQL> update v5 set ename=‟abc‟ where ename=‟SMITH‟;
1 row updated
SQL> update v5 set dname=‟xyz‟ where dname=‟SALES‟;
Error: Cannot modify a column which maps to a non key preserved table.
Generally, in all database systems we cannot perform DML operations through “Complex
Views” to base tables. If we are try to perform DML operations then some table columns
MUARLI SIR 2015
73
are effected and some other table columns are not effected. In Oracle, if we want to
“view” effected, unaffected columns then we are using “user_updatable_columns” data
dictionary.
Eg: SQL> desc user_updatable_columns
SQL> select column_name, updatable from user_updatable_columns where
table_name=‟V5‟;

In Oracle, also we cannot perform DML operations through Complex Views to base table
to overcome this problem Oracle 8.0 introducted instead of triggers in pl/SQL.

TRIGGERS: (PL/SQL concept)

Triggers is also same as “Stored Procedures” and also it will automatically invoked
whenever a DML operations performed on a table.
All database systems having 2 types of triggers .
1. Statement level triggers
2. Row level triggers

In Statement level trigger, triggers body is executed only once per DML statements.
Where as in Row lever trigger, triggers body is executed for each row for DML statements.
Syntax:
Create or replace trigger triggername Trigger specification
before/after Insert/Update/Delete on tablename
[for each row] ----- row level trigger
begin
------ Trigger body
------
------
end;

Difference Between Statement level and Row level Triggers:

Eg: SQL> create table test(col1 date);
Statement level trigger:
SQL> create or replace trigger tv1
after update on emp
begin
insert into test values(sysdate);
end;
/

Testing:
SQL> update emp set sal=sal+100 where deptno=10;
3 rows updated

MUARLI SIR 2015

74
SQL> select * from test;
SQL> delete from test;
SQL> delete trigger tv1;

Row level triggers:

SQL> create or replace trigger tv2
after update on emp
for each row
begin
insert into test values(sysdate);
end;
/

Testing:
SQL> update emp set sal=sal+100 where deptno=10;
3 rows updated
SQL> select * from test;
SQL> drop trigger tv2;
Output:
COL1
------
24-APR-15
24-APR-15
24-APR-15

Row Level Triggers:

In Row level triggers, triggers body is executed per each row for DML statements. That‟s
why we are using “for each row” clause within trigger specification and also DML
transaction “roll back” segment qualifiers. These are old, new.
These qualifiers are used in either in trigger specification or in trigger body when we are
using these qualifiers in trigger body then we must use “:” (column) infront of the
“Qualifier name”.
Syntax:
:old.columnname
Syntax:
:new.columnname

MUARLI SIR 2015

75
Date: 25/4/15

Q) Write a PL/SQL row level trigger on emp table whenever user deleting data then
automatically those deleted records are stored in another table?
Ans: SQL> create table test as select * from emp where 1=2;
SQL> select * from test;
SQL> desc test;
SQL> create or replace trigger tg1
after delete on emp
for each row
begin
insert into test values(:old.empno, :old.ename, :old.job, :old.mgr, :old.hiredate, :old.sal,
:old.comm, :old.deptno);
end;
/
View created
Testing:
SQL> delete from emp where sal>2000;
SQL> select * from test;
10 row selected

Instead of Trigger:
Generally, we cannot perform “DML operations” through complex view to “Base table”. To
overcome this problem Oracle 8.0 introduced “Instead of trigger” in PL/SQL. By default
“Instead of Triggers” are “Row lever Triggers”. “Instead of trigger” are created on “Views”.
Syntax:
Create or replace trigger triggername
instead of insert/update/delete on viewname
for each row
begin
-------
-------
-------
end;

Eg: “Instead of Trigger”:

SQL> create or replace trigger tj1
instead of update on v5
for each row
begin
update dept set dname=:new.dname where dname=:old.dname;
update dept set loc=:new.loc where loc=:old.loc;
end;
/
MUARLI SIR 2015
76
Trigger created
Testing:
Ex 1: SQL> update v5 set dname=‟xyz‟ where dname=‟SALES‟;
Rows updated
Ex 2: SQL> desc user_updatable_columns;
SQL> select COLUMN_NAME, UPDATABLE from user_updatable_columns where
table_name=‟V5‟;
COLUMN_NAMES UPDATABLE
--------------------------------------------
ENAME YES
SAL YES
DNAME YES
LOC YES

SQL> update v5 set loc=‟new york‟ where loc=‟DALLAS‟;

MATERIALIZED VIEWS:
Oracle 8i introduced “Materialized views”. These views are created by database
administrator. “Materialized views” are used in data warehousing applications. Generally,
views doesn‟t store data where as “Materialized views” store data. Generally, “Materialized
views” are used to improve performance of the “Joined” or “Aggregatable Query‟s”.
“Materialized views” stores replication of the remote database into local node.
“Materialized views” stores data same like a table, but whenever when we are refreshing
“Materialized views” it synchronizes data based on “Base Table”.
Syntax:
create materialized view viewname
as select statement;

Difference Between Views, Materialized Views:

VIEWS MATERIALIZED VIEW

1. Views doesn‟t store data 1. Materialized view stores data.
2. Security purpose 2. Improve performance purpose
3. When we are dropping base table 3. If we are dropping “Base table”
also Materialized view can be
accessible.
4. We can perform DML operations 4. We cannot perform DML
on views operations on Materialized views

Before we are creating Materialized view then Database Administrator must give “Create
any Materialized view” privilege to user. Otherwise oracle server returns an error
“Insufficient Privilege.
Syntax:
Grant create any materialized view to username;

MUARLI SIR 2015

77
Eg: SQL> conn scott/tiger
SQL> create materialized view mn1
as
select * from emp;
Error: Insufficient Privileges.
Solution:
SQL> conn sys as sysdba
Enter password: sys
SQL> grant create any materialized view to scott;
SQL> conn scott/tiger
Connected
SQL> create materialized view mn1
as select * from emp;
Materialized View Created.

Date: 27/4/15

In Oracle, one of the materialized view base table must contain a primary key, otherwise
oracle server returns an error.
Eg: SQL> create table test(sno number(10));
SQL> create materialized view mw1
as
select * from test;
Error: Insufficient privileges
SQL> conn sys as sysdba/sys
SQL> grant create any materialized view to scott;
Grant succeeded
SQL> conn scott/tiger
SQL> create materialized view mw1
as select * from test;
Error: table „TEST‟ does not contain a primary key constraint
SQL> create table base(sno number(10) primary key, name varchar2(20));
SQL> insert into base values(&s,‟&n‟);
SQL> select * from base;
SQL> commit;
SQL> create or replace view v1
as select * from base;
SQL> create materialized view mv1
as select * from base;
Here, materialized view also behaves like a view that is in this case materialized view is
also same as “View”. That is whenever we are creating materialized view then
automatically materialized view definitions also permanently stored in database same like
a view definition. In Oracle, if we want to view materialized view definitions then we are
using “user_mviews” data dictionary.
MUARLI SIR 2015
78
SQL> desc user_mviews;
SQL> select query from user_mviews where mview_name=‟MV1‟;
SQL> select rowid, sno, sname from base;
SQL> select rowid, sno, name from v1;
Here, view rowid‟s are also same as base table rowid‟s that‟s why view does not store
data. That‟s why view is also called as “Virtual table” or “Window of a table”. Through the
view only we can view (or) we can access base table data. Generally, view stores stored
query in memory.
SQL> select rowid, sno, name from mv1;
Here, materialized view rowid‟s are different from „Base table‟ rowid‟s are different from
„Base table‟ rowid‟s that‟s why materialized view‟s stores data.
SQL> update based set name=upper(name);
SQL> select * from base;
SQL> select * from v1;
SQL> select * from mv1;
Materialized view also stores data same like a table but, when we are refreshing
materialized view it synchronizes data based on “Base tables”. In Oracle, if we want to
refresh materialized view then we are using “Refresh” procedure from “dbms_mview”
package.
Syntax:
SQL> exec dbms_mview.refresh(„materialized viewname‟);
Dbms_mview: predefined package in PL/SQL
Refresh(„M.V name‟): Predefined procedure name.
SQL> desc dbms_mview;
Eg: SQL> exec dbms_mview.refresh(„mv1‟);
SQL> select * from mv1;

Executive Process:
In Oracle, whenever we are requesting information from the view using „SELECT‟
statement, then Oracle server directly executes view definitions from the „Data
Dictionary‟. That‟s why view does not improves performance.
Whenver, we are creating a materialized view then Oracle server automatically stores
materialized view definitions in a “Data Dictionary” and also result of the query is stored in
different place within database. Whenever user requesting materialized view the select
statement then Oracle server directly retrieve data from materialized view whenever we
are refreshing materialized view then only oracle server executes materialized view
definitions based on these definitions “Base tables” are effected then only Oracle server
synchronize “Base table” data into materialized view. In this process if we are not
refreshing materialized view then “Base table” are never effected. This process
automatically improves performance of the query that‟s why materialized view improves
performance.
Oracle, having two types of materialized views:
1. Complete refresh materialized views
2. Fast refresh materialized views
MUARLI SIR 2015
79
1. COMPLETE REFRESH MATERIALIZED VIEW:
In Oracle, by default materialized views are complete refresh materialized views. These
type of materialized views does not give more performance when we are refreshing
materialized views number of times because in these materialized views whenever we are
refresheing materialized views internally rowid‟s are recreated. If we are not modifying
data in base table also to overcome this problem to improve more performance of query
then oracle introduced “Fast refresh materialized views”.
Syntax:
Create materialized view viewname
refresh complete
as
select statement;
Eg: SQL> select rowid, sno, name from mv1;
SQL> exec dbms_mview.refresh(„mv1‟);
SQL> select rowid, sno, name from mv1;
[here rowid‟s are changed]

Date: 28/4/15

2. FAST REFRESH MATERIALIZED VIEW:

These type of materialized views are also called as Incremental refresh materialized
views. These materialized views performance is very high compare to complete refresh
materialized views because in these materialized views rowids are never changed when
we are refreshing materialized views number of times also.
Syntax:
create materialized view viewname
refresh fast
as
select statement;

Before we are using “Fast Refresh materialized views” then we must capture changes
made in “Base table” for this purpose oracle provided a mechanism called “Materialized
view Log” i.e., when we are creating “Materialized view Log” oracle server capture the
changes from base tables and storing into appropriate data dictionaries. That‟s why before
we are using fast refresh materialized views then we must create materialized view log on
“Base table”.
Syntax:
Create materialized view log on basetablename;
Eg:
SQL> select * from base;
SNO NAME
---------------
1 A
MUARLI SIR 2015
80
2 B
3 C
4 D
SQL> create materialized view log on base;
SQL> create materialized view mk1;
refresh fast
as
select * from base;
SQL> select rowid,sno,name from mk1;
SQL> update base set name=‟xy‟ where sno=1;
SQL> exec dbms_mview.refresh(„mk1‟);
SQL> select rowid, sno, name from mk1;
[Here rowid‟s are not changed only data is affected]

On demand/ On commit:
In Oracle, we can refresh materialized views in 2 ways.
1. Manually
2. Automatically

1. Manually: In manual, method we are refreshing materialized view using

“dbms_mview” package. This method is also called as “On Demand” method. By
default method is “On demand”.
2. Automatically: We can also refresh materialized view without using “dbms_mview”
package. This method is also called as “On Commit” method.
Syntax:
create materialized view viewname
refresh complete/refresh fast on demand/on commit
as
select statement;

Eg:
SQL> create materialized view mk2
refresh fast on commit
as
select * from base;
SQL> select * from mk2;
SNO NAME
---------------------------
1 xy
2 B
3 C
4 D
SQL> update base set name=‟zz‟ where sno=2;
SQL> select * from base;
MUARLI SIR 2015
81
 SNO NAME
---------------------------
1 xy
2 zz
3 C
4 D

SQL> select * from mk2;

SNO NAME
---------------------------
1 xy
2 B
3 C
4 D

SQL> commit;
SQL> select * from mk2;
SNO NAME
---------------------------
1 xy
2 zz
3 C
4 D

In all database systems, if we want to restrict table columns from on user into another
user then also we are using “Views”.

DATA CONTROL LANGUAGE(DCL):

1. Grant -> giving permissions
2. Revoke -> cancel permissions

Creating a User:
SQL> conn sys as sysdba
Enter Password: sys
OR
SQL> conn system/manager

Syntax:
1. Create user username identified by password;
2. Grant connect, resource (or) dba to username;
SQL> conn username/password;

Process for creating a User:

SQL> conn sys as sysdba
MUARLI SIR 2015
82
Enter Password: sys
SQL> create user murali identified by murali;
SQL> grant connect, resource to murali;
SQL> conn murali/murali;
SQL> select * from emp;
Error: table (or) view does not exists;
SQL> conn scott/tiger;
SQL> grant all on emp to murali;
SQL> conn murali/murali;
SQL> select * from emp;
Error: table (or) view does not exists.
SQL> select * from scott.emp;
SQL> create synonym ab for scott.emp;
SQL> select * from ab;
Privilege: Privilege is a right or permission which is used to given to the another “user”.
All database systems having two types of privileges. They are:
1. System Privilege
2. Object Privilege

These Privileges are used to provide security for database objects.

Date : 29/4/15

1. System Privileges:
System Privileges are given by database administrator. These privileges are create table,
create any view, create any materialized view, create session, create procedure, create
any index,….;
Oracle having more than 80 system privileges. These privileges are used to users allowed
to create database object, auditor database objects.
Syntax:
grant system privileges to username1, username2,……;

Eg: SQL> conn sys as sysdba

Enter password: sys
SQL> grant create procedure, create any materialized view, create any index to scott,
murali;

ROLES
Role is nothing but set of system privileges or set of object privilege. User defined rows
are created by database administrator. In a multi user environment, number of users
works on same project. In this case, sometime number of users requires common set of
privileges. In this case only database administrator creating a user defined role and then
assigns set of privileges into role and then only that role given to the number of users.

MUARLI SIR 2015

83
Creating User Defined Role:
Step 1: Create a role
Step 2: Assign system privileges to role
Step 3: Assign rolename to number of users.

STEP 1: Create a Role:

Syntax: create role rolename;
STEP 2: Assign System privileges to Role
Syntax: grant system privilege to rolename;
STEP 3: Assign Rolename to number of users.
Syntax: grant rolename to username1, username2,….;
Eg:
SQL> conn sys as sysdba
Enter password: sys
SQL> create role r1;
Role created
SQL> grant create any materialized view, create trigger, create procedure to r1;
SQL> grant r1 to scott, murali, dinesh;
In Oracle, if we want to view all the system privileges related to role then we are using
“role_sys_privs” data dictionary.
Eg: SQL> desc role_sys_privs;
SQL> select role, privilege from role_sys_privs where role=‟R1‟;
ROLE PRIVILEGE
------------------------------ -------------------------------
R1 CREATE TRIGGER
R1 CREATE ANY MATERIALIZED VIEW
R1 CREATE PROCEDURE

Predefined Roles:
Whenever we are installing oracle server then automatically „3‟ predefined roles are
created. They are:
1. Connect -> end user purpose
2. Resource -> developer purpose
3. DBA -> database administrator purpose

Eg: SQL> desc role_sys_privs;

SQL> select role, privilege from role_sys_privs where role in („connect‟,‟resource‟);

MUARLI SIR 2015

84
NOTE: In Oracle, “connect role” internally having “create session” system privilege. This
privilege is used to users allowed to connect to the “Oracle Server”.
SQL> select role, privilege from role_sys_privs where role in(„DBA‟);

2. Object Privilege:
Object Privileges are given by either database developers or database administrator.
These privileges are used to “Users” are allowed to perform some operations on the
object. Oracle having insert, update, delete, select object privileges related to table.
These „4‟ object privileges also represented by using “all” keyword.
Oracle also supports execute object privilege for PL/SQL objects and also supports read,
write object privileges for “utl_file” package.
Syntax:
grant object privilege on objectname to usernames/rolename/public;
Date: 30/4/15

SQL> conn scott/tiger;

SQL> grant all on emp to murali;
SQL> grant all on emp to r1;
SQL> grant all on emp to public;
NOTE: In Oracle, we can also give object privileges to particular columns. In this case we
must specify number of columns. In this case we must specify number of columns within
parenthesis “()”.
Eg: SQL> grant update(ename) on emp to murali;

 WITH GRANT OPTION:

Who receives “With Grant Option” clause those users allowed to give same object
privileges to another users.
Syntax:
grant object privileges on object name to usernames/public with grant option;

Eg: SQL> grant all on emp to murali with grant option;

NOTE: In all database systems “With Grant Option” clause does not work with roles.
Eg: SQL> grant all on emp to r1 with grant option;
Error: cannot GRANT to a role with GRANT OPTION.
In Oracle, all object privileges are stored under “user_tab_privs” data dictionary.
Eg: SQL> desc user_tab_privs;

REVOKE:
This command is used to “CANCEL” system privileges, object privileges from users.
Syntax:
1. Revoke system privileges from user1,user2,….;
2. Revoke object privileges on object name from username/public;

Eg: SQL> conn sys as sysdba;

MUARLI SIR 2015
85
Enter password: sys
SQL> revoke create any materialized view from scott;
Revoke succeeded
Generally, if we want to restrict table columns from one user into another user then we
are creating views with required columns and then only that view given to the number of
users.
Syntax: grant all on viewname to user1,user2,….;
SQL> conn scott/tiger
SQL> create or replace view vv1
as
select empno, ename, sal, deptno from emp;
SQL> select * from v1;
SQL> grant all on v1 to murali;
SQL> conn murali/murali;
SQL> select * from scott.v1;

Force View (or) Forced View:

In Oracle, we can also create views without using base tables. These type of views are
also called as “ Force Views”.
Syntax:
create or replace force view viewname
as
select * from welcome;
Eg: SQL> create or replace force view v2
as
select * from welcome;
Warning: view created with compilation errors.
SQL> create table welcome (sno number(10));
SQL> alter view v2 compile;
SQL> desc v2;

Read Only views:

In Oracle, we can also create read only views using “with read only” option clause. In
Read only views we cannot perform DML operations. These views are used for reporting
purpose.
Syntax:
Create or replace view viewname
as
select * from tablename where condition with read only;
Eg: SQL> create or replace view v3
as
select * from emp with read only;
SQL> delete from v3 where deptno=10;
Error: cannot delete from view.
MUARLI SIR 2015
86
With Check View:
If we want to provide constraint type mechanism on views then we are using “With Check
Option” clause on views. When a view contains “with check option” clause, then we are
not allowed to insert other than where condition values through view2 base table.
Syntax:
create or replace view viewname
as
select * from tablename where condition with check option;
Eg: SQL> create or replace view v4
as
select * from emp where deptno=10 with check option;
SQL> select * from v4;
Testing:
SQL> insert into v4(empno,ename,deptno) values(1,‟abc‟,30);
Error: view with check option where clause violated.
SQL> insert into v4(empno,ename,deptno) values(1.‟abc,‟10);
1 Row created
SQL> select * from emp;
We can also drop view using “drop view viewname”.

NOTE: In all database systems through the views we can achieve logical data
independence i.e., logical data independence refers whenever we are adding a new entity
in “conceptual level” it is not effected in external level. i.e., whenever we are adding a
new column in a table it is never effected in a view because of Logical Data Independence.
Eg: SQL> alter table emp add address varchar2(10);
SQL> select * from emp;
SQL> select * from v4;

Date : 1/5/15

INDEXES:
Index is an database object which is used to retrieve data very fast from the database.
This process automatically improves performance of query. In all database systems
creating Indexes in 2 ways:
1 Automatically
2 Manually
1. Automatically:
Whenever we are creating “Primary Key” or “Unique Key” then only database servers
automatically creates b-tree indexes on those columns.
2. Manually:
We can also create “Indexes” explicitly by using “create index” command.
MUARLI SIR 2015
87
In Oracle, we are creating index explicitly by using following syntax:
Syntax:
Create index indexname on tablename(columnname);
Generally, indexes are created on “table columns” and also indexes are created by
“database administrator” only.

In Oracle, whenever we are requesting data using “where” clause or “Order by” clause then
only oracle server searching for “Indexes” in appropriate “Data Dictionary” within Oracle
database. If “where” clause columns or “Order by” clause columns having index then Oracle
server uses “Index Scan” method for retrieving data from database. This process
automatically improves performance of the query.
NOTE: In Oracle, whenever “where” clause contains “not equal to” (!=, <>), “is null”, “is not
null” operators then oracle server does not search for indexes. If those columns already
having indexes also.
Oracle having 2 types of Indexes:
1. B-trees Indexes
2. Bit Map Indexes
1. B-Tree Indexes: In Oracle, by default indexes are b-tree indexes. Whenever we are
requesting data using “where” or “Order by” columns, also those columns having b-tree
indexes then Oracle server automatically creates “b-tree structure” based on indexed
columns. In this “b-tree structure” always “Leaf blocks” stores actual data along with
“Rowid‟s”. Based on the “Where” clause columns Oracle server retrieves data from these “Leaf
blocks” using “Index Scan Method”. This process very fastly retrieve data from database.

Syntax:
DEVELOPERs Data Base Administrators
SQL> select * from emp where ename=‟KING‟; SQL> create index in1 on emp(ename);

SQL> create index in1 on emp(ename);

In Oracle, all indexes information stored under “user_indexes” data dictionary.
Eg: SQL> desc user_indexes;
SQL> select index_name, index_type from user_indexes where table_name=‟EMP‟;
Output:
INDEX_NAME INDEX_TYPE
--------------------------------------------
MUARLI SIR 2015
88
PK_EMP NORMAL
IN1 NORMAL

NOTE: In Oracle, if you want to view column names along with Index names then we are
using “user_ind_columns” data dictionary.

Eg: SQL> desc user_ind_columns;

SQL> select index_name, column_name from user_ind_columns where table_name=‟EMP‟;

Output:
INDEX_NAME COLUMN_TYPE
----------------------------------------------
PK_EMP EMPNO
IN1 ENAME

NOTE: In all database systems using indexes we can achieve “Physical Data Independence”
i.e., whenever we are adding indexes(Internal Level) it is not effected in table(Conceptual
Level) But performance is effected.

Date: 2/5/15

In Oracle, if you want to view query performance then database administrator viewing plan
tables. These tables are automatically created by “Oracle server”. Before we are displaying
plan table then we must use “EXPLAIN PLAN FOR” clause in front of the query and then only
we can display plan table using display function from “dbms_xplan” package.
Step 1:
Syntax: SQL> explain plan for select statement;
Step 2: (display plan table)
SQL> select * from table(dbms_xplan.display);
Eg: (without using Indexes)
SQL> select * from emp where ename=‟KING‟;

Testing Performance
SQL> explain plan for select * from emp where ename=‟KING‟;
SQL> select * from table(dbms_xplan.display());

With Using Indexes:

SQL>create index in1 on emp(ename);
SQL> select * from emp where ename=‟KING‟;

Testing Performance
SQL> explain plan for select * from emp where ename=‟KING‟;
SQL> slelect * from table(dbms_xplan.display());

MUARLI SIR 2015

89
Function Based Indexes:
Oracle 8i introduced “Function Based Indexes” by default “function based indexes” are b-tree
indexes.
Generally, whenever we are requesting data using functions (or) expressions then oracle
server does not search for Indexes if those columns already having “Indexes” also. To
overcome this problem Oracle 8i introduced extension of the b-tree indexes called “Function
Based Indexes” which is used to create Indexes on columns along with functions (or)
expressions. Then only Oracle server searching for indexes if we are requesting data using
functions (or) expressions also.

Syntax:
create index indexname on tablename(function name(column name) or expression);

Without using functions based Indexes:

SQL> select * from emp where upper(ename)=‟KING‟;

Testing Performance
SQL> explain plan for select * from emp where upper(ename)=‟KING‟;
SQL> select * from table(dbms_xplan.display());
[here oracle server does not search for indexes]

With using function based Indexes:

SQL> create index in2 on emp(upper(ename));
SQL> select * from emp where upper(ename)=‟KING‟;

Testing Performance
SQL> explain plan for select * from where upper(ename)=‟KING‟;
SQL> select * from table(dbms_xplan.display());

SQL> desc user_indexes;

SQL> select index_name, index_type from user_indexes where table_name=‟EMP‟;
INDEX_NAME INDEX_TYPE
----------------------------------------------
IN2 FUNCTION BASED NORMAL

VIRTUAL COLUMNS
Oracle 11g introduced “Virtual Columns” in a table which stores “Stored Expression” directly in
Oracle database prior (mean before) to 11g Database Administrators only stores “Stored
Expressions” indirectly using “VIEWS” , “Function Based Indexes”. Where as when we are
using “Virtual Columns” we can store “Stored Expressions” directly in database using
“Generated Always As” clause.

Syntax:
columnname datatype(size) generated always as(stored expression)[virtual]
MUARLI SIR 2015
90
Eg:
SQL> create table test(a number(10), b number(10), c number(10) generated always as
(a+b) virtual);
SQL> insert into test(a,b) values(20,40);
SQL> select * from test;

A B C
-----------------
20 40 60
If we want to view “Virtual Column” expression then we are using “data_default” property
from “user_tab_columns” data dictionary.
SQL> desc user_tab_columns;
SQL> select column_name, data_default from user_tab_columns where table_name=‟TEST‟;
COLUMN_NAME DATA_DEFAULT
-------------------------------------------------
C A+B

Oracle having two types of b-tree indexes.

1. Non unique b-tree Indexes
2. Unique b-tree Indexes

By default, all automatic Indexes are “Unique b-tree Indexes” and also all explicit Indexes are
“Non unique b-tree Indexes”. By default, “Unique b-tree Indexes” performance is very high
compare to “Non unique b-tree Indexes”. We can also create “Unique b-tree Indexes”
explicitly by using following syntax.

Syntax:
Create unique index indexname on tablename(columnname);
Eg: SQL> create unique index in3 on emp(ename);
Index created
NOTE: We are not allowed to create unique indexes on duplicate value columns.
SQL> create unique index in4 on emp(job);
ERROR: cannot CREATE UNIQUE INDEX: duplicate key found.

2. BIT MAP INDEX:

Oracle 7.3, introduced Bit Map Indexes. Bit Map Indexes are used in Data Ware housing
applications. Generally, Bit Map Indexes are created on “Low Cardinality Columns”.
Syntax:
create bit map index indexname on tablename(columnname);

Date: 4/5/15

MUARLI SIR 2015

91
Eg 1: Cardinality of empno = 14/14 = 1 (high cardinality) -> b-tree
Eg 2: Cardinality of job = 5/14 = 0.367…. (low cardinality) -> bit map indexes
Whenever we are creating a “Bit Map Index” then oracle server automatically creates a “Bit
Map Table” based the column values in a table. Whenever user requesting data using “Where”
clause then oracle server automatically uses “Index Scan” internally on “Bit Map Table” also
perform operations on “Bits” within “Bit Map Table”. When we are using logical operators in
“Where” clause then oracle server automatically converts the resultant Bit Map into “ROWID”
using a “Internal Bit Map” function.

SQL> create bitmap index abc on emp(job);

SQL> desc user_indexes;
SQL> select index_type, index_name from user_indexes where table_name=‟EMP‟;
Output:
INDEX_TYPE INDEX_NAME
---------------------------------------
NORMAL PK_EMP
BITMAP ABC
In Oracle, we can also drop Index using
“drop index indexname”;

SYNONYMS

Synonyms is a database object which hides another schema user name, object name.
Synonyms also provides security for the original object. It is an “Alias name” or “Reference
Name” for the original objects. Synonyms are created by database administrators.
All database systems having 2 types of synonyms
1. Public Synonyms
2. Private Synonyms

In all database systems by default Synonyms are “PRIVATE Synonyms”.

Syntax:
MUARLI SIR 2015
92
create synonym synonymname for
username. Objectname@database link;

In Oracle, if you want to create Public Synonyms then we are using “CREATE PUBLIC
SYNONYM” privilege to user otherwise oracle server returns as “Insufficient Privilege Error”.
Syntax: grant create public synonym to username;
Syntax:
create public synonym synonymname for
username. Objectname@database link;

Scott/tiger SQL> conn murali/murali SQL> conn a1/a1

SQL> grant all on empSQL> select * from scott.emp; SQL> select * from scott.emp;
to murali, a1; SQL>create synonym abc for scott.emp; SQL> select * from abc;
SQL> select * from abc;
SQL> create public synonym xy for
scott.emp;
Error: Insufficient Privilege
SQL> conn sys as sysdba;
Enter password: sys
SQL> grant create public synonym to
murali;
SQL> conn murali/murali;
SQL> create public synonym xy for
scott.emp;
SQL> select * from xy; SQL> select * from xy;

All Synonyms information stored under “ user_synonyms” data dictionary.

SQL> desc user_synonyms;
We can also drop synonyms by using
“ drop synonym synonymname”

LOCKS

Locking is a mechanism which prevents unauthorized access for our resource. Whole database
systems having 2 types of locks.
1. Row Level Locks.
2. Table Level Locks

1. Row Level Locks: Oracle 6.0 introduced “Row Level Locks”. If you want to establishes
“Row Level Locks” then we are using “ for update” clause in all databases. This clause used
only in “Select” statement. In this method we are locking set of rows.
Syntax: select * from tablename where condition for update[no wait];
Whenever we are performing locks then another user query the data query but they cannot
perform DML operation whenever we are using “Commit” and “Rollback” locks are released.

MUARLI SIR 2015

93
 Date: 5/5/15

Scott/tiger Murali/murali
SQL> select * from emp where deptno=10 SQL> update scott.emp set sal=sal+100
for update; where deptno=10;
[we cannot perform DML operations]
SQL> commit;[for releasing lock]
NO WAIT: “No wait” is an optional clause used along with “for update” clause. When we are
using “no wait” clause oracle server automatically returns control to the current session if
another user not releasing locks also. In this case oracle server returns an error.
ORA-0054: resource busy

Scott/tiger Murali/murali
SQL> select * from scott.emp where
deptno=10 for update;
SQL> select * from emp where deptno=10
for update no wait;
ORA-0054: resource busy
SQL>_

DEFAULT LOCKS: In all database systems whenever we are using DML operations then
automatically database servers internally uses “Exclusive Locks”. These are also called as
“Default Locks”.
Eg:
Scott/tiger Murali/murali
SQL> update emp set sal=sal+100 where
deptno=10; SQL> update scott.emp set sal=sal+100
4 rows updated where deptno=10;
5 SQL> commit;[for releasing locks] [we could not perform DML operations
because of “Default Locks”]

DEAD LOCKS:
When we try to perform same DML operations on same resource at a time through different
sessions for a “single user” or through different users then oracle server automatically returns
an error.
ORA-0060: dead lock detected

Scott/tiger Scott/tiger
SQL> update emp set sal=sal+100 where SQL> update emp set sal=sal+100
deptno=100; where depnto=20;
3 rows updated 6 rows updated
SQL> update emp set sal=sal+100 7
where deptno=20;
ORA-0060: dead lock detected
SQL> commit; [for releasing locks]

MUARLI SIR 2015

94
TABLE LEVEL LOCKS:
In this method we are locking a table. Oracle having 2 types of table level locks. “Table Level
Locks” are handled by Database Administrators.
1. Share lock
2. Exclusive lock

1. Share lock: When we are using this lock another user query the data but they cannot
perform DML operations and also at a time number of users can lock the resource.
Syntax: lock table tablename in share mode;
Here, also when ever we are using “Commit” or “Rollback” then only automatically locks are
released.
Eg:
Scott/tiger Murali/murali
SQL> lock table emp in share mode; SQL> select * from scott.emp;
SQL> lock table scott.emp in share mode;
SQL> update scott.emp set sal=sal+100;
[we cannot perform DML operations]
SQL> commit; [for releasing lock]

2. Exclusive lock: When we are using “exclusive locks” then another user query the data but
they cannot perform DML operations and also at a time only one user lock the resource.
Syntax: lock table tablename in exclusive mode;
Eg:
Scott/tiger Murali/muarli
SQL> lock table emp in exclusive mode; SQL> select * from scott.emp;
SQL> lock table scott.emp in share mode;
SQL> commit;[for releasing locks]

NOTE: in all database systems whenever we are using cursor locking mechanism then
automatically database servers internally uses “Exclusive mode”.

SEQUENCE

Sequence is an database object which is used to generate “Sequence numbers” automatically.

Generally, sequences are used to generates “Primary key values” automatically. Generally,
sequences are created by database administrator. Sequence is an “Independent database
object”. Once a sequence has been created then number of users simultaneously accepts that
“sequence”.
Syntax:
create sequence sequencename
start with n
increment by n
minvalue n
maxvalue n

MUARLI SIR 2015

95
cycle/nocycle
cache/nocache;

In Oracle, if we want to access sequence values then we are using 2 pseudo columns.
1. Currval
2. Nextval

Syntax (currval): sequencename.currval;

Syntax (nextval): sequencename.nextval;
These Pseudo columns are used in Insert, Update, Delete, Select statements.
In Oracle, if we want to generate sequence values using “SELECT statement” then we must
use “dual” table.
Syntax: select sequencename.currval from dual;
Syntax: select sequencename.nextval from dual;

“Currval pseudo” column returns current value of the sequence if sequence session already
having a value. Where as nextval pseudo column returns next values for sequence update.
In Oracle, if we want to generate first sequence number then we must use nextval pseudo
column. Because currval pseudo column returns value if sequence session already having a
value.
In all database systems, we are using “Currval” pseudo column after using “Nextval” pseudo
columns only.
Eg: SQL> create sequence s1
start with 5
increment by 2
maxvalue 100;

Sequence created
SQL> select s1 currval from dual;
Error: sequence s1.currval is not yet defined in this session
SQL> select s1.nextval from dual;
5
SQL> select s1.nextval from dual;
7
SQL> select s1.nextval from dual;
9
SQL> select s1.nextval from dual;
11
SQL> select s1.currval from dual;
11

MUARLI SIR 2015

96
Date: 6/5/15

NOTE: In sequence we can also change “Sequence parameter values” by using “ALTER”
command.
Syntax:
alter sequence sequencename parametername newvalue;
Eg: alter sequence s1 increment by -1;
Sequence altered
In all database systems we can modify all sequence parameter values except “Start with”
clause.
Eg: SQL> alter sequence s1 start with 3;
Error: cannot alter starting sequence number

NOTE: Generally, start with values cannot be less than “min value”
Eg: SQL> create sequence s1
start with 4
increment by 1
minvalue 5
max value 100;
Error: START WITH cannot be less than MIN VALUE

CACHE: Cache is an “Optional” clause defined in “Sequence” object which is used to access
“Sequence values” very fastly. This process automatically improves performance of the
application. Generally, cache is an memory area which is used to store set of “Sequence
Numbers”.
Generally, whenever we are creating a “Sequence” those sequences are created in hard disk.
Whenever we are requesting sequences using a client tool, then server process checks
requested sequence is available in cache memory area. If it is not available then control goes
to hard disk and returns “Sequence value” from the disk and storing into “Cache memory
area”. Then only that value return to client tool. This default process retrieves “Sequence
Values” slowly from the database because each and every time server process interacting with
disk. To overcome this problem to access sequence values very fastly then database
administrator defines “Cache parameter” in sequence object. Based on the parameter value
cache memory pre-allocates set of “Sequence numbers”. Whenever we are requesting
sequences then server process, directly access sequence values from the “Cache memory” are
not from the “Hard disk”. This process automatically improves performance of the application
because this process reduces Disk I/O(input/output). Whenever system crashes (or) power
failure then automatically cache values are lost.
In Oracle, cache default value is “20” and also cache minvalue is “2”.
Eg: SQL> create sequence s1
start with 1
cache 1;
Error: The number of values to CACHE must be greater than 1.
Generally, sequences are used to generate “Primary key” values automatically.
MUARLI SIR 2015
97
Eg: SQL> create sequence s1
start with 1;
SQL> create table test(sno number(10) primary key, name varchar2(10));
SQL> insert into test(sno,name) values(s1.nextval, „&name‟);
Enter value for name: murali
SQL> /
Enter value for name: abc
SQL> /
Enter value for name: xyz
SQL> select * from test;
Output:
SNO NAME
------------------------
1 MURALI
2 ABC
3 XYZ

In Oracle, all sequences information stored under “user_sequences” data dictionary.

Eg: SQL> desc user_sequences;
We can also drop a sequence using “drop sequence sequencename”.
Eg: SQL> drop sequence s1;

MERGE
Oracle 9i introduced “Merge” statement. Merge is an DML command which is used to transfer
data from “source table” into “target table” when table structure are same.
Generally, merge statement used in data warehousing applications. In merge statement we
are using update, insert commands. This command is also called as “UPSERT” command
(UP-> update, SERT->insert).
Syntax:
Merge into targettablename
using sourcetablename
on(join condition)
when matched then
update set targettablecol1= sourcetablecol1,…….
when not matched then
insert(targettablecolumnnames) values(sourcetablecolumnnames);

Date: 8/5/15

Eg: SQL> select * from dept; (target table)

SQL> create table depts as select * from dept;
SQL> insert into depts values(1,‟a‟,‟b‟);
SQL> select * from depts;

MUARLI SIR 2015

98
SQL> merge into dept d
using depts s
on(d.deptno=s.deptno)
when matched then
update set d.dname=s.dname, d.loc=s.loc
when not matched then
insert (d.deptno,d.dname,d.loc) values(s.deptno, s.dname, s.loc)
5 rows merged
SQL> select * from dept;
Through “Merge” statement we cannot update “ON” clause columns.
SET OPERATORS:
Set operators are used to retrieve data from single (or) multiple tables. These operators are
also called as “Vertical Joins”.
1. Union -> It returns unique values and also automatically sorting data.
2. Union all -> It returns unique + duplicate data. (no automatic sorting)
3. Intersect -> It returns common values
4. Minus -> Values are in first query those values are not in second query.

Eg:
SQL> select job from emp where deptno=10
union
select job from emp where deptno=20;
NOTE: Whenever we are using set operators, corresponding expressions must belongs to
same data type.
Eg:
SQL> select deptno from emp
union
select dname from dept;
Error: expression must have same data type as corresponding expression
NOTE: Always set operators returns first query column names (or) alias names as “Column
headings”.
Eg:
SQL> select dname from dept
union
select ename from emp;
Output:
DNAME
--------
Accounting
ADAMS
……….
NOTE: In Oracle, in corresponding expressions not belongs to same data type also then we
are retrieving data from multiple query‟s using “Set Operators”. In this case we must use
appropriate “Type conversion” function.
MUARLI SIR 2015
99
Eg:
SQL> select deptno from emp
union
select dname from dept;
Error: expression must have same data type as corresponding expression
SOLUTION:
SQL> select deptno “deptno”, to_char(null) “dept name” from emp
union
select to_number(null),dname from dept;
Output:
DEPTNO DNAME
-------------------------
10
20
30
ACCOUNTING
SALESMAN
………………
CONVERSIONS:
Converting one data type into another data type is called “Conversions”.
Oracle having 2 types of conversions.
1. Implicit conversions
2. Explicit conversions

1. Implicit conversions:
A. In Orcale, when an expression contains string representing pure number (eg: „1234‟) then
oracle server automatically converts “String type” in “Number type”.
Eg: SQL> select sal+‟100‟ from emp;
It works
B. In Oracle, whenever we are passing number into character functions then oracle server
automatically converts number type into character type.
Eg: SQL> select length(1234) from dual;
Output: 4
C. In Oracle, whenever we are passing date string into date functions then oracle server
automatically converts “Data String” into “Date Type”. But here, passed parameter must
be in “Default Date Format”.
Eg: SQL> select last_day(‟12-aug-05‟) from dual;
Output: 31-aug-05.

MUARLI SIR 2015

100
Date: 9/5/15

IMPLICIT CONVERSION TABLE

FROM TO ASSIGNMENT EXPRESSION
EVALUATION
Varchar2 (or) char Number Yes Yes
Varchar2 (or) char Date Yes Yes
Number Varchar2 (or) char Yes No
Date Varchar2 (or) char Yes No

EXPLICIT CONVERSIONS:
Using “Explicit conversions” functions we can also convert one data type into another data
type explicitly.
DECODE():
Decode() is an conversion function which is used to decoding the values.
Decode() is also same as “if…. then … else if” constructor in PL/SQL. Decode() internally used
equality operator (=).
Syntax:
Decode (columnname, value1, statement1,value2, statement2, statement);
Eg: SQL> select ename, sal, deptno, decode(deptno, 10,‟ten‟, 20, „twenty‟, others) from emp;
Output:
ENAME SAL DEPTNO DECODE
---------------------------------------------------------
10 TEN
20 TWENTY
30 OTHERS
Using “decode()” we are decoding “number” into string (or) “Single character” into string.
NOTE: If we want to modify data conditionally within a table using “ SQL” then we must use
“Decode()” function.
Eg:
Q) Write a Query to update “comm.” of the emp‟s in emp table based on following condition:
1. If job= „CLERK‟ then update comm into 10% of sal.
2. If job= „SALESMAN‟ then update comm into 20% of sal.
3. If job= „ANALYST‟ then update comm is 30% of sal
4. Else comm is 40% of sal.
Ans: SQL> update emp set comm= (decode(job, „CLERK‟, sal*0.1, „SALESMAN‟, sal*0.2,
„ANALYST‟, sal*0.3, sal*0.4));
In Oracle, if we want to display aggregate functional values in tabular format and also if we
want to display those values conditionally using PIVOTING(rows are converted into columns).
Then we must use “Decode()” with in “Group By” clause.
Eg: SQL> select job, sum(decode(deptno, 10, sal)) “deptno 10”,
sum (decode(deptno,20,sal)) “deptno 20”,
sum (decode(deptno,30,sal)) “depnto 30” from emp group by job;
Output:

MUARLI SIR 2015

101
JOB DEPTNO 10 DEPTNO 20 DEPTNO 30
-------------------------------------------------------------------------
ANALYST 6400
CLERK 1700 2700 1350
MANAGER 2650 3175 3050
PRESIDENT 5200
SALESMAN 5800
Eg: SQL> select dname, sum(decode(job, „CLERK‟, 1,0)) “CLERKS”
sum (decode(job, „SALESMAN‟,1,0)) “salesman”
sum (decode(job, „ANALYST‟, 1,0)) “analyst”
from emp e, dept d
where e.deptno=d.deptno
group by dname;
Output:
DNAME CLERKS SALESMAN ANALYST
-------------------------------------------------------
ACCOUTING 1 0 0
RESEARCH 2 0 2
SALES 1 4 0
Case Statement:
Case statement also used to “Decoding” the values. “Case Statement” performance is very
high compare to decode conversion function. Oracle 8.0, introduce “Case statement” and also
“Oracle 8i” introduced case conditional statement. This is also called as “ Searched Case
Statement”.
NOTE: Decode internally used “Equality operator” (=), where as in “Case statement” we can
also use all SQL operator (<,>,<=,>=,<>,Like, is ,etc……) explicitly.
METHOD 1:
Syntax:
case columnname
when value1 then statement1
when value2 then statement2
---------------------------------
……………………………………………..
else statements end;
Eg: SQL> select ename, sal, deptno, case deptno
when 10 then „ten‟
when 20 then „twenty‟
else „others‟ end
from emp;
OUTPUT:
ENAME SAL DEPTNO DEPTNO DECODE

METHOD 2: (Case Coditional Statement (8i))

Syntax:
MUARLI SIR 2015
102
case
when columncondition1 then statement1
when columncondition2 then statement2
…………………………………………………………………
…………………………………………………………………
else statements end;
Eg: SQL> select ename, sal,
Case
when sal<1000 then „Low Salary‟
when sal between 1000 and 3000 then „Medium Salary‟
when sal in(3100,3200,3500,3800) then „Special Salary‟
else „Other Salary‟ end
from emp;
Date: 11/5/15

Oracle also having following explicit conversion functions.

1. to_number()
2. to_char()
3. to_date()

1. to_number(): It is used to converting a “String” representing numeric value with format

into numeric value without format.
Eg: SQL> select „$45.6‟+3 from dual;
Error: invalid number
SQL> select „45.6‟+3 from dual;
O/P: 48.6
SQL> select to_number(„$45.6‟)+3 from dual;
Error: Invalid number

NOTE: When ever we are using “to_number()” and also if we try to convert “String”
representing numeric value with format into numeric value without format then we must use
“second parameter” as same as “first parameter” format. Then only oracle server
automatically converts “String type” to “Number type”.

MUARLI SIR 2015

103
Solution:
SQL> select to_number(„$45.6‟, „$99.9‟)+3 from dual;
O/P: 48.6

Not for other character:

Eg: SQL> select to_number(„a45.6‟, „a99.9‟) +3 from dual;
Error: Invalid number format model

2. to_char(): to_char() is an “Over Loading” function which is used to convert “Number

type” into “Character type” and also used to convert “Date type” into “Character type” or
“Date String ”.
Converting “Number type” into “Character type”:
Syntax: to_char(number, „character format‟)

FORMAT ELEMENTS:
9 -> representing Number
g -> Group separator (,)
d -> Decimal Indicator (.)
L -> Local Currency
$ -> Dollar ($)

Group Separator (,) and Decimal Indicator (.):

Eg: SQL> select to_char(1234567, „99g99g999‟) from dual;
O/P: 12,34,567
OR
SQL> select to_char(1234567, ‟99,99,999‟) from dual;
O/P: 12,34,567

Combination of both g & d:

SQL> select to_char(1234567, „99g99g999d99‟) from dual;
O/P: 12,34,567.00
OR
SQL> select to_char(1234567, ‟99,99,999.99‟) from dual;
O/P: 12,34,567.00

Leading Zero:
Eg: SQL> select to_char(123, ‟0999‟) from dual;
O/P: 0123
SQL> select to_char(123, „9990)) from dual;
O/P: 123
SQL> select to_char(123, „$999‟) from dual;
O/P: $123

MUARLI SIR 2015

104
Local Currency:
Eg: SQL> select to_char(123, „L999‟) from dual;
O/P: $123

NOTE: If we want to display our own currency then we are using “nls_currency” parameter in
third parameter of to “to_char()” function. In this case this parameter must be specified within
single quotes. Before we are using this currency then we must use “Local Currency(L)” in the
format model of the second parameter.
Eg: SQL> select ename, to_char(sal, „L99g99g999d99‟, „nls_currency = IndRs‟) from emp;
Output:
ENAME SAL
----------------------
SMITH INDRS 800.00
ALLEN INDRS 1,400.00

Eg: SQL> select ename, nvl(mgr, „no manager‟) from emp;

Error: Invalid manager
Solution:
SQL> select ename, nvl(to_char(mgr), „no manager‟) from emp;
Output:
ENAME TO_CHAR(MGR), „NO MANAGER‟
---------------------------------------------------
KING NO MANAGER

Eg: SQL> select to_char(sysdate, „DD/MM/YYYY‟) from dual;

O/P: 11/05/2015

3. to_date(): It is used to convert “Date String” into “Date Type”.

Eg: SQL> select „15/AUG/05‟+3 from dual;

Error: Invalid Number
Solution:
SQL> select to_date(„15/AUG/05‟)+3 from dual;
Output: 18/AUG/05

NORMALIZATION

Normalization is an “Scientific Process” which is used to decomposing a table into number of

tables. Normalization process automatically avoids insertion, updation, deletion problems and
also this process reduces duplicate data.
In design phase of the SDLC, database designers designs “Logical Model” of the database. In
this logical model only they are using “Normalization Process” through “Normal Forms”.

MUARLI SIR 2015

105
In 1970‟s, E.F.CODD written a paper “Relational Model Data For Large Shared Data Banks”. In
this paper only E.F.CODD introduced first three “NORMAL FORMS”.

NORMAL FORMS:
1. First Normal Form
2. Second Normal Form
3. Third Normal Form
4. BCNF
5. Fourth Normal Form
6. Fifth Normal Form

Date: 12/5/15

1. FIRST NORMAL FORM(1NF):

If a table is in 1NF, in that table each column contain a “Atomic value” and also identifying
record “Uniquely” using a “Key”.

Process: Identifying a “Repeating Groups” and put it into separate table in more “Atomic
form”. By default, 1NF process a table is an “Child table”. Because in this table one column
having duplicate data.

2. SECOND NORMAL FORM(2NF):

If a table is in 1NF and also all “ Non key attributes” are fully functionally dependent on “Total
Candidate Key”. Always 1NF deals with “Atomicity” where as “2NF” deals with relationship
between “Key and Non Key attributes”.

MUARLI SIR 2015

106
If a table is in 1NF and also that table contains partial “Non Key attributes” then that table not
in “2NF”.
Process: Identifying partial non key attributes which depends on “Partial Key attributes” put
into separate table. This table is called “2NF” table. By default this table is an “Master table”.
In this table only all “Non Key attributes” are “Fully Functionally dependent on total candidate
key”.

Date: 13/5/15

MUARLI SIR 2015

107
Phase 1:
1. Ecode ---(<---FD----)--- Hours(Worng) E101-- 4,9
2. Projcode----(<---FD---)- Hours(Wrong) P1- 4,5,7
3. Ecode+Projcode --- Hours(Correct)
F -> Fully Functional Dependent

Phase 2:
1. Ecode----(<---FD---)-- Dept(Correct) E101-- Systems
2. Projcode-----(<---FD---)- Dept(Wrong) P1-- systems, sales, admin
Once partial dependencies is satisfied we want to go for “Fully Functional Dependent”.

Phase 3:
1. Ecode----(<---FD---)-- Depthead(Correct)
MUARLI SIR 2015
108
2. Projcode-- Depthead(Wrong)

Date: 14/5/15

Before Normalization process in the above research having Insertion, Updation, Deletion
problems.
Insertion Problem: In the above resource table when we are trying to insert particular
department employee then we must assign project code. If project code is not available then
we need to supply null values for the project code field. This is called “Insertion Problem”.
Updation Problem: In the above resouce table ecode, dept, depthead attribute values are
repeated. Whenever resource table having large amount of data and also when we are
modifying data then we must change these three values correctly. Otherwise, Inconsistency
Problem occurs. This is called “ Updation Problem”.
Deletion Problem: In the above resource table when we try to delete a employee record
then automatically deparment details also deleted. This is called “Deletion Problem”.
Whenever we are using “Normalization Process” then automatically Insertion, Updation,
Deletion problems avoided.
Without using Normalization Process

Data Redundancy Problems

1. Disc space wasted
2. Data Inconsistency
3. DML Operations become slow

MUARLI SIR 2015

109
A Table is in “1NF”:
1. Data in each column should be atomic:
Is there is no multiple values saperated by column (,).
2. Table does not contain repeating column groups.
3. Identifying a record uniquely using a “Primary Key”.

1.

2.
3. SOLUTION:

Date: 15/5/15

THIRD NORMAL FORM(3NF):

If a table is an 2NF and also all non key attributes are only dependent on “Candidate key”.
If a table is in 2NF and also “Non Key Attributes” which dependents on another “Non Key
Attributes” then that table not in “3NF”.
Process: Identifying a “Non Key attributes” which depends on another “Non Key attributes”
but into saperate table. This table is called “3NF” table. By default this table is an “Master
table”. In this table only all “Non key attributes” are only dependent on “Candidate Key”.

MUARLI SIR 2015

110
LOGICAL DIAGRAM:

DATE: 16/5/15

1NF -> Remove Repeating Groups

MUARLI SIR 2015

111
2NF -> Remove Pratial Attributes
3NF -> Remove Attributes which are not dependent on Candidate Key.
(OR)
Remove Non-Key Attribute which are dependent on another Non-Key Attribute.

CLUSTERS

Cluster is an “Database Object” which contains group of tables together and it will “Shares
same Data Blocks”.
Generally, clusters are used to improves performance only.
In all database clusters tables must have a “Common Column Name”. This common columns
is also called as “Cluster Key”.
In all databases, whenever we are submitting “Inner Join” (or) “Outer Join” then database
servers checks “FROM” clause tables are available in clusters (or) not.
If those tables are available in cluster then database servers retrieve data very fastly from
the “Cluster Tables”.

Creating Clusters:
Step 1: Create a Cluster
Step 2: Create an Index on Cluster
Step 3: Create Cluster Tables.

Step 1: Create a Cluster:

Clusters are created based on common column. This common column is also called as
“Cluster Key”.
MUARLI SIR 2015
112
Syntax: create cluster clustername(common colname datatype(size));

Step 2: Create an Index on Cluster:

Syntax: create index indexname on cluster clustername;

Step 3: Create Cluster Tables:

Syntax:
create tablename(commoncolname datatype(size), col1 datatype(size),….)
cluster
clustername(commoncolname);

Eg:
Sql> create cluster emp_dept(deptno number(10));
Sql> create index in1 on cluster emp_dept;
Sql> create table emp1(empno number(10), ename varchar2(10), sal number(10), deptno
number(10)) cluster emp_dept(deptno);
Sql> create table dept1(deptno number(10), dname varchar2(10), loc varchar2(10)) cluster
emp_dept(deptno);
Sql> desc emp1;
Sql> desc dept1;
Sql> insert into emp1 values(1, „abc‟, 2000, 10);
Sql> insert into dept1 values(10, „xyz‟, „hyd‟);
Sql> select * from emp1;
Sql> select * from dept1;
NOTE: In all database system cluster tables having some “Rowid‟s”.
Eg: sql> select rowid from emp;
Sql> select rowid from dep1;
Generally, we cannot drop cluster if cluster having tables. To overcome this problem Oracle 8i
introduced to drop cluster along with tables using “Inclusing tables” clause.
Syntax:
Sql> drop cluster clustername including tables;
Eg: sql> drop emp_dept including tables;
Cluster dropped;

Date: 18/5/15

In Oracle all Clusters information stord under “user_clusters” data dictionary.

In all relational databases using Clusters we can achieve physical data independence, because
whenever we are adding clusters table structure does not effects but only performance is
effected when we are using “joins”.

Candidate Keys:
1. Empno+skill
2. Ename+skill
MUARLI SIR 2015
113
3. Voterid+skill

SUPER KEY: A column (or) combination of columns which uniquely indentifying a record in a
table is called “Super Key”.
CANDIDATE KEY: A minimal “Super Key” which uniquely identifying a record in a table is
called “Candidate Key”
(or)
A “Super Key” which does not contains another “Super Key” is also called as “Candidate Key”.

FUNCTIONAL DEPENDENCY(FD):
If any given two tuples in a relation “r” then if X(an attribute set) agrees then Y(another
attribute set) cannot disagree then only X->Y is called “Functional Dependency”.
X->Y here Y is Functionally Dependent on X
X functionally determines Y. (here X is also called a “Determinent” and Y is also called as
“Dependent”)

BCNF: When a table is in BCNF then every determinant is a “Candidate Key”.

When a table contains multiple composite candidate keys and also those candidate keys are
overlapped and also one candidate key non key attribuite which depends on another
candidate key non key attribute then only we are using BCNF process.

Generally, when a table contains multiple composite candidate keys then deletion problem
occurred to overcome this problem database designers uses “Boyce Code Normal Form” in
database design.

Generally, second and third normal forms deals with relationship between key and non key
attributes. Where as BCNF deals with relationship between non key attributes within
composite candidate keys itself.

PROCESS: Identify one composite candidate key non key attributes which depends on
another composite candidate key non key attributes put into saperate table. This table is also
called as “BCNF table”. In this table only every determinant behaves like a candidate key and
also in this BCNF table. There like a candidate key and also in this BCNF table. There is no
non key attributes.

MUARLI SIR 2015

114
NOTE: If a table contains single composite candidate key then BCNF is also same as 3NF.

Date: 19/5/15

MULTIVALUED DEPENDENCY:
In a table, if a determinant determines multiple values in another column within same table is
called “Multivalued dependency”. Multivalued dependency represented by using Double Arrow
marks (“->->”).
 Single Value Multiple Columns

FOURTH NORMAL FORM(4NF): If a table is in 4NF then that table doesnot contain more
than one independent multivalued attributes.
If a table contain minimum three(3) attributes and also using combination all these attributes
uniquely identification of the record and also one attribute set of values which depends on
another attribute set of values and also same attribute set of values which depends on
another attribute set values and also some attributes are not logically related then only we
are using 4NF process.

PROCESS: Identify independent multivalued attributes put it into saperate tables. These
tables are called 4NF tables these tables does not contain more than one independent
multivalued attributes.

MUARLI SIR 2015

115
Generally, when a table contains more than one independent multivalued attribute. Then
those tables having more duplicate data to overcome this problem for reducing duplicate data
then database designers uses 4NF process in this process one table does not contain more
than one independent multivalued attribute.
ASSUMPTION: Each employee has multiple projects and also each employee having multiple
skills.

In the above resource table whenever an employee got a new project then we have to supply
null values for the skill attribute and also whenever an employee got new skill then we have
to supply null values for the „project id‟ attribute. These null value problems are automatically
avoided when we are using 4NF process.

FIFTH NORMAL FORM(5NF):

If a relation cannot be decomposed into further relations is called 5NF i.e., if a table is in 5NF
the table doesnot contains cyclic dependencies.
Generally, if a table contains more than two attributes and also identifying a record using
combination of all these attributes and also all these attributes set of values related to one
another then only we are using 5NF.

NOTE: When a table contains multivalued attributes and also some attributes are not logically
related when only we are using 4th NF process where as when a table contains multivalued
attribute and also all attributes are logically related then only we are using 5th NF process.

5TH Normal Form is also called as “Projection Joined Normal Form” because in 5th NF if
possible then we can decomposing a table into number of tables but when we are joining
those tables then resultant record must be available in resource table.

Eg:

ORACLE SERVER ARCHITECTURE

MUARLI SIR 2015

116
Oracle server mainly consist of two parts:
1. Storage Area
2. Instance
1. Storage Area: Whenever we are installing oracle server automatically three physical files
are created in harddisk these files are called as Storage Area (or) Physical database.
1. Data files(.DBF)
2. Control files(.ctl)
3. Redolog files(.log)
All databases having two types of structures
1. Logical Structure
2. Physical Structure

LOGICAL STRUCTURE: A structure which is not visible in operating system is called “Logical
Structure”.
-> Logical structure contains database objects like table, views, synonyms, clusters,
sequences.
-> Logical structure is handled by either developers (or) database administrators.

PHYSICAL STRUCTURE: A structure which is visible in operating system is called Physical

structure. Physical structure contains datafiels, control files, redolog files.
-> Physical structure is handled by database administrator only.

Date: 20/5/15

Data Files: Data files stores database objects physically that is these files stores tables,
views, sequences, indexes permanently into hard disk these files extension is “.dbf”.
-> In Oracle all datafiles information stored under “dba_data_files” data dictionary.
Eg: sql> conn sys as dba
Enter password: sys
MUARLI SIR 2015
117
Sql> desc dba_data_files;
Sql> select file_name from dba_data_files;
CONTROL FILES: Control files controls all other files within storage area these files
extension is “.ctl”.
-> Control files also stroes physical database information these fiels are used by database
administrator in backup and recovery process.
-> In Oracle, all control files information stored under “v$control file” data dictionary.
Eg: sql> desc v$ control file;
Sql> select name from v$control file;

REDOLOG FILES: Redolog files stores commited data from the redolog buffer. Redolog files
also used by database administrator in backup and recovery process.
-> In Oracle, all log files information stored under “v$logfile” data dictionary.
Eg: sql> desc v$logfile;
Sql> select member from v$logfile;

INSTANCE: Whenever we are performing any operations within database those all operations
also first performed in Physical memory area. This Physical Memory area is also called as
“Instance”. This Instance is available in RAM memory area this instance having two parts.
1. SGA
2. Processes
SGA: SGA is also called as System Global Area (or) Shared Global Area. SGA memory area
consist of set of buffers these are
1. Database buffer cache
2. Shared pool
3. Java pool
4. Large pool
5. Redolog buffer

-> Whenever we are submitting sql, pl/sql code that code is stored in library cache within
shared pool Library code always reduces parsing. Library cache also stores cache values
defined in Sequence database object.
-> Shared Pool also contains dictionary cache which executes “dcl” related objects and also
checks username and passwords given by the client connect to the server.
-> Whenever we are requesting a table using a client tool then server process always checks
requested table is available in database buffer cahce. If requested table is not available in
buffer cache then dbwriter process checks requested table is available in data files. If
requested table is available in data files then copy of the table transferred into database
buffer cache. Then only server process fetching that table information from database buffer
cache into client tool.
-> Java Pool executes java related objects and also large pool performs large number of
operations and also redolog buffer stored new information for the transaction.

MUARLI SIR 2015

118
UNDO: In Oracle, when we are performing dml transactions then new values for the
transaction is automatically stored in dirty buffer within database buffer cache and also old
value for the transaction is automatically stored in undo area within buffer cahce. Whenever
user using commit. Then automatically new value for the transaction is stored in data files
within storage area and also old values for the transaction is automatically stored in UNDO
file within storage area.
-> In Oracle 9i, onwards we can also retrieve accidental data after commiting the transaction
using flashback queries these flashback queries internally uses UNDO file.

Date: 21/5/15

REDO: Whenever we are performing transactions new value for the transaction is
automatically storing “Redolog” buffer. Whenever “User” using “Commit (or) 1/3 fill of
Redolog buffer” the Redolog buffer data automatically transferred into Redolog files. These
Redolog files used by database administrator in backups or recovery process.

MUARLI SIR 2015

119
PROCESSESS: Oracle Instance internally having following background processes. These are:
1. Dbwr(database writer)
2. Lgwr(log writer)
3. Ckptr(check pointer)
4. Pmon(process monitor)
5. Smon(system monitor)

1. DBWR: Dbwr process fetches data from data base buffer cache and store that data in
data files within storage area.
2. LGWR: Lgwr process fetches data from Redolog buffer and that data stored into “Redolog
files”.
3. CKPTR: Whenever data base writer perform some task then check pointer automatically
generates an unique identification number for every transaction. This number is
automatically stored in data files, control files, Redolog files. This number is also called as
System Change Number(SCN). This number is used by data base administrator in “Flash
Back Querys”. Flash Back querys retrieves accidental data based on a specific point of
time. After commiting the trasactions only in “Flash back queries” we are using as of
clauses.
Syntax:
Sql> select * from tablename as of scn system change number;
We can also view system change number using “current_scn” property from “v$database”
data dictionary in sys as dba user only.
Eg:
Sql> conn sys as sysdba
Enter the Password: sys
Sql> create table b1(sno number(10);
Sql> desc v$database;
Sql> select current_scn from v$database;
Output:
CURRENT_SCN
----------------
1012963
MUARLI SIR 2015
120
 Sql> insert into b1 values(50);
Sql> commit;
Sql> select count(*) from b1;
Output: 1
Sql> select count(*) from b1 as of scn 1012963;
Output: 0
4. PMON: Process monitor cleans up user processes if users does not disconnect properly
from the server.
5. SMON: System monitor process recovery all processes. When ever problems are occurred
in an process. That‟s why system monitor is also called as “Instance Recovery”.
PGA: (Private Global Area)
In Oracle, server process contains a memory area which uniquely indentified by each client
connect to the server. This memory area is also called as “Private Global Area”. In pl/sql all
connections are executed within PGA memory area.
Hierarchical Query‟s:
In Relational Databases we can also stores hierarchical data. If you want to store hierarchical
data then minimum three columns required in a table. And also in those three columns tow
columns must belongs to same data type. And also having some relation these two columns
are also called as “Hierarchical Columns”.
If you want to retrieve hierarchical data from a relational table then we are using following
clauses.
1. Level
2. Start with
3. Connect by

1. LEVEL: Level is a Pseudo column which automatically assigns numbers to level in a tree
structure. Levels are also starting with Level number “1” (one).
2. START WITH: Using “Start with” clause specifies starting condition with in tree structure.

Date: 22/5/15

Syntax: start with condition

3. CONNECT BY: Using “Connect BY” clause we are specifying relationship between
hierarchical columns using “Prior” operator.
Syntax: connect by prior parent columnname=childcolumnname;
Syntax:
select level, columnname from tablename
where condition
start with condition
connect by prior parentcolname=childcolname;

Q) Write an hierarchical query to display the employees who are working under other
employees route node onwards from emp table?
ANS:
MUARLI SIR 2015
121
sql> select level, ename from emp
Start with mgr is null
Connect by prior empno=mgr;
OUTPUT:
LEVEL ENAME
------ ----------
1 KING
2 JONES
3 SCOTT
4 ADAMS
3 FORD
4 SMITH
2 BLAKE
3 ALLEN
3 WARD
3 MARTIN
3 TURNER
3 JAMES
2 CLARK
3 MILLER

NOTE: Oracle 9i, introduced “sys_connect_by_path()” function which returns path of the
hierarchy in tree structure. This function accepts two parameters.
Syntax:
Sys_connect_by_path(colname,‟delimiter name‟);
Eg:
sql> select level, sys_connect_by_path(ename,‟->‟)
from emp
start with mgr is null
connect by prior empno=mgr;
Q) Write a hierarchical query to display the employees who are working under “BLAKE” from
emp table?
ANS:
sql> select level, sys_connect_by_path(ename,‟->‟)
from emp
MUARLI SIR 2015
122
start with ename=‟BLAKE‟
connect by prior empno=mgr;
OUTPUT:
LEVEL ENAME
------------------
1->BLAKE
2->BLAKE->ALLEN
2->BLAKE->WARD
2->BLAKE->MARTIN
2->BLAKE->TURNER
Eg: sql> select level, sys_connect_by_path(ename,‟->‟)
from emp
start with ename=‟BLAKE‟
connect by empno= prior mgr;
OUTPUT:
LEVEL NAME
----------------
1-> BLAKE
2->BLAKE->KING

PRIOR: Prior is an “Unary Operator” used along with “Connect By” clause. When we are
using prior operator infront of the parent column (mgr) then oracle server uses bottom to up
search within tree structure where as when we are using prior operator infront of the child
column then oracle server uses top to bottom search within tree structure.
EXECUTION: Based on the “Start with” condition oracle server take a “Prior” operator
column value and that value is searching in another hierarchical column. Then only
corresponding column data retrieve from the table.

Q) Create a hierarchical table based on the following tree structure?

ANS:
Sql>create table test(ename varchar2(10), empno number(10), mgr number(10));
Sql> insert into test values(„anand‟,6000,8000);
Sql> select * from test;
OUTPUT:
ENAME EMPNO MGR
------------------------------------
ANAND 6000 8000
MILLER 6001 8000
ANIL 8000 9000
JAMES 8001 9000
SMITH 9000 10000
JONES 9001 10000
MURALI 10000 NULL

MUARLI SIR 2015

123
Mainly, three conditions to build hierarchical table:
1. Table must contain 3 columns
2. One column must have a “NULL” value(for root node)
3. Any two columns must be same data type and interrelated (i.e., one column all values
must be in another column for example “mgr” column is parent column and “empno”
column is child column because parent column consist of “NULL” value)

NOTE: Whenever we are using “Order By” clause in hierarchical query‟s database servers
sorting the data but changes the hierarchical structure within tree structure. To overcome this
problem Oracle introduced “Order Siblings By” clause within hierarchical query‟s.
Syntax:
order siblings by columnname[asc/desc]
Eg:
Sql> select level, ename
from emp
start with mgr is null
connect by prior empno=mgr
order siblings by ename;

NOTE: Oracle 10g, introduced “connect_by_root” operator which returns root node in the
hierarchy.
Syntax: connect_by_root columnname;
Eg:
Sql>select level, ename, connect_by_root ename
from emp
start with mgr is null
connect by prior empno=mgr
order siblings by ename;

Date: 23/5/15

TRANSACTION CONTROL LANGUAGE(TCL):

Transction: A logical unit of work in between two points is called “Transaction”.
All database systems having 2 transactional command.
MUARLI SIR 2015
124
1. Commit
2. Savepoint

1. COMMIT: This command is used to save the transaction permanently to database.

Syntax: commit;
2. SAVEPOINT: A logical mark between transactions is called “Savepoint”.
Syntax: savepoint savepointname;
3. ROLLBACK: This command is used to “UNDO” the transactions from memory.
Syntax: rollback;
ROLLBACK TO PERTICULAR SAVEPOINT:
Syntax: rollback to savepointname;
Eg: Sql> insert into emp(empno) values(1);
Sql> update emp set sal=sal+100 where ename=‟SMITH‟;
Sql> savepoint s1;
Sql> insert into emp(empno) values(2);
Sql> update emp set sal=sal+100 where ename=‟KING‟;
Sql> savepoint s2;
Sql> delete from emp where empno=2;
Sql> rollback to s1;
Sql> commit;
Sql> select * from emp;

NESTED TABLE:
Oracle 8.0, introduced nested table. A table within another table is also called as “Nested
Table”. Basically, nested table is an user defined datatype which is used to store number of
data items in a single unit.
Generally, if we want to store number of data items in a single unit then we are using Index
by table in PL/SQL. But these tables are not allowed to store permanently into database. To
overcome this problem Oracle 8.0 introduced extension of the Index by tables called “Nested
Table” which stores number of data items and also these tables are allowed to store
permanently into oracle database using SQL.

CREATING NESTED TABLE:

Step 1: create an Object type.
Step 2: create an Nested table type
Step 3: create an Actual table.

Step 1: Create an Object Type:

Object Type: Object type is an user defined type which stores different data types into single
unit. It is also same as “Structures” in “C language”.
Syntax:
create or replace type typename
as object (attribute1 datatype(size), attribute2 datatype(size),…….);
Step 2: Create an Nested Table Type:
MUARLI SIR 2015
125
Create an Nested table type by using Object type.
Syntax:
Create or replace type typename
as table of objecttypename;
Step 3: Create an Actual Table:
Syntax:
create table tablename(col1 datatype(size),….,coln nestedtabletype)
nested table coln store as anyname;

Eg: sql> create or replace type obj1

as object(bookid number(10), bookname varchar2(10), price number(10));
/
Sql> create or replace type
abc as table of obj;
/
Sql> create table student(sno number(10), sname varchar2(10), col3 abc) nested table col3
store as zzzz;
Table Created.
Sql> desc student;
NAME TYPE
----------------------
SNO NUMBER(10)
SNAME VARCHAR2(10)
COL3 abc

If we want to store data into nested table type then we must use constructor. Here
constructor name is also same as type name.

INSERTION:
Sql> insert into student values(1,‟murali‟,abc(obj1(101,‟java‟,900), obj1(102,‟plsql‟,500)));
1 row created.
Sql> select * from student;

We can also retrieve modify delete nested table data saperately by using table operator i.e.,
in place of table name we are using following syntax in select, update, delete statements.
Syntax:
table (select nested tablecolname from relational tablename);

MUARLI SIR 2015

126
Eg: sql> select * from table(select col3 from student);

OUTPUT:
BOOKID BOOKNAME PRICE
--------------------------------------
101 JAVA 900
102 PLSQL 500.

Eg: sql> update table(select col3 from student) set price=1000 where bookid=101;
OUPUT:
BOOKID BOOKNAME PRICE
-------------------------------------------
101 JAVA 1000

Date: 25/5/15

PARTITIONS

Partitions: At the time of table creation a table can be decomposed into number of partions is
called “Partition Table”. Partition Tables are used in dataware housing applications. Partition
Tables are created by database administrator and used in back up and recovery process.
Partition Tables are used to improve performance of the applications in backup and recovery
process. Partition Tables are created on very large data bases. Partition Tables are created
based on key column. This is called “Partition Key”.
Oracle having 3 types of partitions.
1. Range partition
2. List partition
3. Hash partition

1. Range Partition: In this partition we are partitioning tables based on range of values.
Syntax:
create table tablename(col1 datatype(size), col2 datatype(size),….)
partition by range(key colname)(partition partitionname values less than(value),…..,partition
partitionname values less than(max value));

If we want to view particular portion then we are using following syntax:

Syntax:
select * from tablename partition(partitionname1, partitionname2,……);

Eg:
sql> create table test(sno number(10), name varchar2(10), sal number(10))
partition by range(sal)(partition p1 values less than(1000), partition p2 values less
than(2000), partition p3 values less than(max value));

MUARLI SIR 2015

127
Table created
Sql> insert into test values(&no, „&name‟, &sal);
Sql> select * from test partition(p1);
OUTPUT:
SNO NAME SAL
-----------------------------------
1 ABC 500

2. List Partition: Oracle 9i, introduced list partition in this partition we are partitioning
tables based on list of values.
Syntax:
create table tablename(col1 datatype(size), col2 datatype(size),…..)
partition by list(key colname)
(partition partitionname values(value1, value2,…),………………………………….. partition
partitionname values(default));
NOTE: Generally, if we want to create partition based on character datatype column then we
are using “List Partition”.
Eg:
Sql> create table test(sno number(10), name varchar2(10)) partition by list(name)
(partition p1 values(„India‟,‟Nepal‟), partition p2 values(„us‟,‟uk‟,‟canada‟)
partition p3 values(default));
sql> insert into test values(&sno,‟&name‟);
sql> select * from test;
OUTPUT:
SNO NAME
-----------------------
1 AUS
2 JAPAN
3 INDIA

Sql> select * from test partition(p1);

OUTPUT:
SNO NAME
----------------------
1 AUS
2 JAPAN

3. HASH PARTITION:
In “Hash Partition” oracle server only creates partitions by using “Hash Algorithm”.
Syntax:
create table tablename(col1 datatype(size), col2 datatype(size),……)
partition by hash(keycolname) partition anynumber;

Eg:
MUARLI SIR 2015
128
Sql> create table test(sno number(10), sal number(10)) partition by hash(sal) partition 5;
Table created
Sql> insert into values test(&sno, &sal);
If we want to view partitions then we are using “user_tab_partitions” data dictionary.
Eg:
Sql> desc user_tab_partitions;
Sql> select partition_name from user_tab_partitions where table_name=‟TEST‟;

END OF SQL(STRUCTURED QUERY LANGUAGE(SEQUEL))

By DINESH LINGALA

MUARLI SIR 2015

129