PL/SQL Pl/SQL woks both in the Oracle Sever(through stored procedures, stored functions, database triggers and

packages), and in the Oracle Development Environment(Oracle Forms, Oracle Reports, and Oracle Graphics) It supports SQL Data types. PL SQL bridges the gap and the need for procedural programming capabilities. Benefits of PL/SQL : Improved performance : It is used to group SQl statements together within a single block and to send the entire block to the server in a single call, thereby reducing the network traffic. Modularised program Development : Group logically related statements within blocks. Nest sub-blocks inside larger blocks to build a powerful program Break down complex problems into a set of manageable, well-defined , logical modules Place reusable PL/SQL code in libraries to be shared between Oracle applications or store it in an Oracle server to make it available to any userapplication. Portability : PL/SQL programs can run anywhere the Oracle server is installed, you do not need to tailor them to each new environment. Identifiers : It allows you to declare variables, cursors, constants, and exceptions and then use them in SQL and Procedural statements. It allows to declare variable belonging to scalar, reference, composite and LOB data types Declare variable based on the columns in the tables.

Procedural language Control Structures : Execute a sequence of statements conditionally Execute a sequence of statements iteratively in a loop Process the individual rows returned by a multi-row query with an explicit cursors. Error Handling : Process Oracle-server errors with exception handling routines Declare user-defined error conditions and process them with exception-handling routines Structure of a PL/SQL Block Declare(optional) Contains all the variables, constants, cursors and user-defined exceptions that will be used in the executable section Begin(Mandatory)-----Executable section Contains SQL statements to manipulate data in the database and PL/SQl statements to manipulate data in the data-block Exception(Optional) Contains actions to be performed when errors and abnormal conditions arise in the executable section End; set serveroutput on for dbms_output.put_line to give output in SQL

declare vname students.studname%type; vtotal; begin select studname into vname from students where rollno=1; select total into vtotal from students where rollno=1; dbms_output.put_line('Name of Student is :'||vname); dbms_output.put_line('Total of Student is :'||vtotal); end; /

PL/SQL Block Types Anonymous Blocks: These are unnamed blocks. They are declared in an application where they need to be executed and are passed to the Server for execution at run-time. Example : Triggers in Oracle Developer consists of such blocks [Declare] Begin [Exception] End; Sub-Programs : These are named PL/SQL blocks that can accept parameters and can be invoked as and when required. They are declared as Procedures or Functions. Generally use procedures to perform an action and a function to compute and return a value. Procedure name is Begin --statements [Exception] End; Function name return datatype Is Begin --statements return value; [Exception] End;

Instead of hardcoding the data type and size of a variable. Example . …. RECORD: A record is a group of related data items stored as fields. variable type_name. Salary constant number : = 10000 PL/SQL variables 1) Scalar These are the data types that correspond to the data types of columns in tables. TRUE .). each with it’s own name and data type. Boolean data types(only stores NULL. salary. Type type_name IS RECORD ( field declaration. RECORD. Example : vsalary employees. hire date. field declaration. Must contain one or more components of any scalar type. FALSE) vflag Boolean.Declaring PL/SQl variables Declare Birthdate date. NESTED TABLE and VARRAY. allows groups of fields to be defined and manipulated in PL/SQL blocks.salary%type vgrade number(7. vvgrade vgrade %type:=10.2). vvflag Boolean :=TRUE 2) Composite Composite data types such as records. you can use the %type attribute to declare a variable according to another previously declared variable or a column of a table. you have different kinds of data about an employee like name. PL/SQL also supports Boolean datatype Character. Deptno number(2) NOT NULL : =10. This data is dissimilar in type but logically related. Numeric Date and Boolean types of data types. Also known as Collections. The %type attribute. . They are of 4 types : TABLE. They have internal components that can be handled and manipulated individually. Location varchar2(20) : = ‘Mumbai’.

. which stores the INDEX BY Table elements. The number and the data types of the underlying database column may change at run time. An INDEX BY Table : Is similar to an array Must contain 2 components A Primary Key of data type Binary_Integer that indexes the INDEX BY TABLE A Column of a scalar or record data type. They use a primary key to provide you with array-like access to rows. TABLE : Objects of TABLE type are called INDEX By TABLES.first_name:=’Ramesh’. This attribute is useful when retrieving a row into variables with the Select * From tablename. Declaring Records with %rowtype The %rowtype attribute is used to declare a record based on a collection of columns in a database table or view Example : Declare Emp_record employees%rowtype.%rowtype INDEX BY BINARY_INTEGER. Can increase dynamically because it is unconstrained. last_name varchar2(10)). Emp_record. identifier type_name.field_name. Type type_name is Table of column_type/variable %type/table.Type emp_data IS RECORD (first_name varchar2(10). emp_record emp_data.column%type/table. To reference or initialize an individual field : Record_name. The number and data types of the underlying database columns need not be known.

Declare Type ename_table_type IS TABLE OF employees%rowtype INDEX BY BINARY_INTEGER.Returns the index number that succeeds the index n in a table TRIM-removes oen element from the end of the PL SQL Table TRIM(n)-removes n elements from the end of the PLSQL Table DELETE. Type ename_table_type IS TABLE OF employees.n)-removes all elements in the range m…n from a PL SQL table. it contains no keys or no values. . Ename_table ename_table_type.Type ename_table_type IS TABLE OF employees. Initially such a table is not populated.Returns the index number that preceeds the index n in a table Next(n). Birthdate_table birthdate_table_type. My_emp_table emp_table_type. LAST. Ename_table ename_table_type.removes all elements from PL/SQL table DELETE(n)-removes the nth element from the PLSQL table. Following methods are used by the INDEX BY Tables : Exists(n)-returns TRUE if the nth element exists Count-returns number of elements that a PL/SQL table currently contains FIRST.first_name%type INDEX BY BINARY_INTEGER. Ename_table(1) :=’Ramesh’.returns the first and last index numbers Prior(n). An explicit executable statement is required to initialize (populate) the Index By Table. DELETE(m.first_name%type INDEX BY BINARY_INTEGER. Type birthdate_table_type IS TABLE OF DATE INDEX BY BINARY_INTEGER. Example : Set server output on. Birthdate_table(1) :=sysdate-10.

Conditional Constructs IF Statement If-then. .first_name). For i in my_emp_table.. Else(optional) Statement 3. / 3) Reference These are the pointers to other program items. Statement 2. End if.. my_emp_table. Begin For i in 100 . End. Statement 2. End if. End loop. v_count Loop Select * into my_emp_table(i) from employees where employee_id=i.V_count number(3):=104. If-then-else-end if. If <condition(s)> then Statement 1. Statement 4. If <condition(s)> then Statement 1.last Loop Dbms_output. End loop.(Not a part of the syllabus) 4) LOBs Also called locators that specify the location of large objects(such as graphic images that are stored out of line.end if.first .put_line(my_emp_table(i).

select total into vtotal from students where rollno=1. if total of Rollno=1 is more than 275. end. end if. vtotal students.studname%type. Statement 8.put_line(Vname||' is a very good student'). select total into vtotal from students where rollno=1. vtotal students. Else(optional) Statement 7. else display "Average student" declare vname PL/SQL Block to display "Very Good Student". begin select studname into vname from students where rollno=1. Statement 4. "Average student" if total between 250 and <275. If <condition(s)> then Statement 1.If-then-elsif-elsif-else-end if. begin select studname into vname from students where rollno=1. End if.put_line(Vname||' is an average student'). if vtotal>=250 then dbms_output. if total of Rollno=1 is more than 275. Elsif <condition(s)> then Statement 3. Statement 2. / PL/SQL Block to display "Very Good Student". Statement 6. and "poor student" if total less than 250 declare vname students. Elsif <condition(s)> then Statement 5. . else dbms_output.

End.if vtotal>=275 then dbms_output. Begin Vappraisal := CASE vgrade When ‘A’ then ‘Excellent’ When ‘B’ then ‘Very Good’ When ‘C’ then ‘Good’ Else ‘No such grade’ End.put_line(Vname||' is a very good student'). end if.put_line(‘Grade ‘||vgrade||’ Appraisal ‘ || vappraisal). Declare vgrade char(1):=upper(‘&p_grade). vappraisal varchar2(20). end.put_line(Vname||' is an average student'). Dbms_output. / Nesting of Ifs allowed as per situation. / . CASE Expressions Case selector When expression1 then result1 When expression2 then result2 When expression3 then result3 When expressionN then resultN Else resultN+1(optional) End. Set serveroutput on. elsif vtotal>=250 and vtotal<275 then dbms_output. else dbms_output.put_line(Vname||' is a poor student').

end.20 loop if mod(x.2)=0 then dbms_output. end. begin for num in reverse 1. statement2.put_line('Number '||x||' is Odd'). begin for num in 1. else dbms_output..highvalue loop statement1. b number:=0. a:=a+1. / Print odd and even numbers from 1 to 20 and their count declare a number:=0. end.10 loop dbms_output. b:=b+1. end if..put_line('There are '||b||' odd numbers'). end loop.. begin For x in 1. / . / Numbers 1 to 10 in reverse order declare num number(2).put_line('There are '||a||' even numbers').put_line('number '|| num).LOOPS 1) For <variable> in [reverse] lowvalue.put_line('number '|| num).10 loop dbms_output. dbms_output. end loop. end loop Numbers 1 to 10 in order declare num number(2).put_line('Number '||x||' is Even').. dbms_output. end loop.

begin while num<=100 loop dbms_output.15. begin loop dbms_output. end loop. end loop. exit when <condition>..20.. / . end. num:=num+5.. print 5.10. statement2.put_line(num). num:=num+10.put_line(num)..100 declare num number:=10.50 declare num number:=5..30...While <condition> loop statement1. print 10. end.. end loop. exit when num>50. statement2. end loop. / loop statement1.20.

Following are the steps followed for handling explicit cursors : 1)declare Here we declare the cursor by naming it and define the structure of the query to be performed within it. 4)close It releases the active . and manipulated in the block’s executable section. Close Cursor_name . Loop Fetch cursor_name into variable1.CURSORS Cursors are names given to private SQL work areas created by the Oracle server to execute SQL statements and to store the processing information. EXPLICIT For queries that return more than one row. There are 2 types of cursors : IMPLICIT These are declaredc by PLSQL implicitly for all DML and PLSQL Select statements. 3)Fetch and process Here you actually fetch the rows into variable if required and carry out all the business logic or processing. The row returned by the query are called the active set and are now available for fetching Open Cursor_name. explicit cursors are declared and named by the programmer. 2)Open This statement executes the query. including queries that return only one row. Explicit Cursors Used to process each row returned by a multiple-row Select statement. Declare Cursor cursor_name is Select ……. Exit When End loop. It is now possible to reopen the cursor to re-establis a fresh active set. variable 2.

dbms_output.studname.put_line('Name :'||vname).Do not use the INTO clause in the cursor declaration. If the row need to be processed in a particular order. loop fetch cstudent into vrollno. / Here. end. last_name from employees. Emp_record emp_cursor%rowtype. end loop. vtotal students. . begin Open cstudent. total of all the students declare cursor cstudent is Select Rollno. vname. Begin Open emp_cursor.rollno%type. "notfound" is a cursor attribute which is used to find whether the cursor currently contains a record or has reached the end of the file or recordset Explicit Cursor attributes %isopen – Evaluates to True if the cursor is open %notfound. or else the last record will be displayed twice. vtotal. vname to print the name.Evaluates to TRUE if the most recent fetch returns a row %rowcount-Evaluates to the number of rows returned so far Example to retrieve row from employees table and populate them to another table with a similar structure Declare Cursor emp_cursor is select empid. you may use the order by clause in the from students. dbms_output. close cstudent. vrollno students.put_line('Roll number :'||vrollno). exit when cstudent%notfound should be done immediately after fetch. exit when cstudent%notfound. dbms_output.Evaluates to TRUE if the most recent fetch does not return a row %found.put_line('Total :'||vtotal).

--implicit close and implicit loop exit End. Commit. Statement2.first_name). End. . Thus you can open and close an explicit cursor several times in a block. End loop.deptno=80 then Dbms_output. because the cursor is opened. / Cursor For Loop It is a shortcut for processing explicit cursors. Begin For emp_record IN emp_cursor --implicit open and fetch occurs Loop If emp_record. deptno from employees. For each execution. Retrieving the employees one by one and printing them if they belong to department 80. and the cursor is closed automatically.empid. End loop. name) values(emp_record. Declare Cursor emp_cursor is select first_name .last_name) End loop.Loop Fetch emp_cursor into emp_record. Each time you open the cursor with a different set of values. returning a different active set. Set serveroutput on. End if. emp_record.put_line(‘Employee ‘ ||emp_record. Insert into temp_emp(empno. rows are fetched once for each iteration in the loop. the previous cursor is closed and reopnened with a new set of values. / Cursors with Parameters You can pass parameters to the cursors. Close emp_cursor. For record_name IN Cursor_name Loop Statement1. Exit when emp_cursor%notfound. the loop exits when the last row is processed.

Only that table’s rows will be locked whose column is referred to in the For update [of column] clause. ‘Sales_Rep’). : ‘ || emp_record.employee_id ||’ Employee name :’ || emp_record. Close emp_cursor Open emp_cursor(60. last_name from employees whre department_id=p_deptno and job_id=p_job. Declare cursor emp_cursor(p_deptno number. End loop.last_name).In such a case.] Also. and each of there formal parameters will have to be passed an ACTUAL parameter or value in the Open statement. Hence it is important that no other user /transaction modifies those values.In the cursor declaration.last_name). you can lock those row retuned by the cursor through the FOR Update clause.. You have to specify the Formal parameter data types. you need to specify the FORMAL parameters. but not the size. but it is recommended by Oracle for better readability and maintenance. [It is not necessary that the for update clause refers to a column . Begin Open emp_cursor(80. To prevent any other user/transaction from changing those values. For emp_record IN emp_cursor Loop Dbms_output.put_line(‘Employee no. : ‘ || emp_record. Sometimes it may happen that the rows which you want to lock are already locked by another user/transaction. ‘Programmer’).employee_id ||’ Employee name :’ || emp_record. since the update/delete which you want to perform is based on those values. For emp_record IN emp_cursor Loop Dbms_output. because of the for update . when querying multiple tables with a join. Close emp_cursor End. End loop. you can use the For update clause to ensure row locking to particular table of the join. / For Update [Of column] clause Sometimes you may want to perform some update or delete based on the values that the cursor has returned to you.put_line(‘Employee no. p_job varchar2) is select employee_id.

But for this you need to lock the rows with the FOR UPDATE clause so that the rows are locked on open.salary*1.deprtment_id=60 FOR UPDATE OF SALARY NOWAIT. Where current Of clause Sometimes you may want to update or delete the row being currently addressed. . To avoid this type of waiting.department_id and employees. count(*) as staff from employees group by department_id) t2 where t1. employee_id.department_id.department_id=t2.department_id=60 FOR Update of salary NOWAIT. the active set will contain the department number. (Select department_id..staff from departments t1. This is possible through the WHERE CURRENT OF clause.clause your block will keep on waiting till the rows are released by the other user/transaction. Begin For emp_record in sal_cursor Loop If emp_record.salary<5000 then Update employees set salary=emp_record. without the need to refer to the ROWID. you can use the NOWAIT clause with the For update clause.10 WHERE CURRENT OF sal_cursor. department name. / Cursors with subqueries Declare Cursor my_cursor is Select t1.department_name. departments where employees. departments d where d.department_id and d. salary from employees e. End.department_id=departments. such that there are more than 3 employees working in each such department. so that instead of waiting indefinitely. t1. Declare Cursor sal_cursor is Select e.. End loop.department_id and t2. and the total number of employees working in that department. End if. last_name.department_id=e. last_name. Declare Cursor emp_cursor is Select employee_id.staff>=3 When this cursor is opened. the control is immediately passed to your program so that it can do some other work before trying to acquire the lock. t2.department_id. department_name from employees.

Declare them in the declaration section. and oracle server will raise them implicitly Explicitly raised : User-defined : A situation which is abnormal from the user point of view. Declare them in the declaration section and raise them explicitly Declare Begin Exception When Exception1 [or exception2…. The exception being raised may be user-defined or pre-defined. When an exception is raised. you can specify an exception handler to perform some other actions. then the control is not passed back to the enclosing block or environment or point where the exception was raised. Do not declare them and oracle server will raise them implicitly No_data_found Too_many_rows Inavlid_cursor Zero_divide Dup_val_on_index. the control is passed to the exception handler in the exception section. then the PL SQL block terminates with failure and the exception is propogated to the calling environment.] then Statement1. Exceptions are raised in 2 ways : An oracle-defined error occurs and the appropriate exception is raised automatically. Statement2. In such cases.EXCEPTIONS An exception is an identifier in a PL SQL Block that is raised during the execution of a block that terminates it’s main body of actions.. If the exception is successfully handled. etc Non Predefined Oracle Server : Other Oracle server errors. Exception types Implicitly raised : Predefined Oracle Server : 20 most common errors. But.. You explicitly raise an exception by issuing the RAISE statement within the block. . If an exception is raised and there is no corresponding exception handler. but the PL SQl block terminates successfully.

besides the Exception section Trapping Predefined Exceptions : declare vtotal students. . unless you use the OTHERS exception handler. since the data is not found. OTHERS is the last exception handler that is defined optionally Guidelines for Trapping Exceptions : Begin the Exception-handling section with the EXCEPTION Keyword Define the required number of exception-handlers. others are not trapped. dbms_output. which traps any exception which is not yet handled. Statement6. each with it’s own set of actions When an exception occurs PL SQL processes only one Exception-handler before leaving the block Place the OTHERS clause after all the other exception-handlers There can be only one OTHERS clause Exceptions cannot appear anywhere else. / The exception handling section traps only those exceptions that are specified. since when the error is encountered.] then Statement3. the control is not passed to the immediate next line. Hence.] End. begin select total into vtotal from students where rollno=16.put_line('the roll number is not correct').total%type. and the line which is after the line that caused the error is not executed. [When others then Statement5. Statement4.When Exception3 [or exception4…. but to the exception handler. an error has occurred and the control is transferred to the exception handler. exception when no_data_found then dbms_output.put_line('Check the roll number'). end.. / Here.

Define p_deptno=20 Declare e_emps_existing EXCEPTION. Such errors can be trapped by declaring it first . Exception When e_emps_existing then Dbms_output. After doing this . End. but are not associated with name(exception name/error name). BEGIN Delete from departments where department_id=&p_deptno. -2292). / Trapping Non-Predefined Exceptions : There are certain non-predefined standard Oracle errors. / Functions for trapping Exceptions : SQLCODE-returns the numeric value for the error code SQLERRM-Returns the message associated with the error number ‘ || . Commit. you may refer to that exception by name and write a specific handler for it. PRAGMA directs the PL SQL compiler to associate all occurrences of that Oracle error number with the given exception name within that block. PRAGMA EXCEPTION_INIT(e_emps_existing.declare vtotal students.put_line(‘Cant delete department to_char(&p_deptno) ||’ It has got employees existing ‘).put_line('More than 1 row selected'). Example : Oracle returns error number –2292 when you try to delete a parent record which has got some child records. which always return a standard Oracle server error number. The PRAGMA EXCEPTION_INIT utility tells the compiler to associate an exception name with an Oracle error number. end. They are raised implicitly. exception when too_many_rows then begin select total into vtotal from students where total>100. or by using the OTHERS handler.

/ Trapping User-Defined Exceptions : Declare the exception in the declarative section Raise the exception in the executable section Handle the exception in the Exception-handling section declare low_mark exception. v_error_message varchar2(200).english%type. v_error_message).put_line('Failed in English'). End. Begin Exception When OTHERS then Rollback.SQLCODE value 0 1 +100 negative number Description No Exception encountered User-defined exception No_data_found exception Some standard Oracle server error number Example : to store all the error-numbers and error-messages in an ERRORS table Declare v_error_code number. venglish students. end. insert into errors values(v_error_code. / . v_error_message :=SQLERRM. exception when low_mark then dbms_output. end if. if venglish<35 then raise low_mark. begin select english into venglish from students where rollno=2. v_error_code:=SQLCODE.

If SQL%NOTFOUND then Raise e_invalid_department. End.put_line(‘No such department number’). End if. Begin Update departments set department_name=’&p_department_desc’ where department_id=&p_department_number. End if. / Here no_data_found cannot be used . Begin For erec in emp_cursor loop Begin Select ……. Commit. If SQL%NOTFOUND then Raise e_no_rows. Sub-blocks can handle an exception or pass the exception to the enclosing block Declare … … e_no_rows exception. When e_no_rows then……. / .if marks in english are less than 35 then the exception is raised and the message is displayed. -2292). since it is an update statement and not a select statement. pragma exception_init(e_integrity. Exception When e_invalid_department then Dbms_output. End. End loop. e_integrity exception. Define p_department_desc =’Planning n esearch’ Define p_department_number=20 Declare e_invalid_department EXCEPTION. End. Exception When e_integrity then…….. Update…….

specified number for exceptions between -20000 to -20999 user-specified message upto 2048 bytes long. Raise_application_error can be used in either(or both) the executable section and the exception section. It also handles the unhandled exceptions. nonpredefined or userdefined error .Raise_Application_Error Procedure : This procedure is used to communicate a predefined exception/error with a nonstandard error number and a user-friendly error message from stored subprograms. Raise_application_error(error_number. The error number and message are displayed to the user just like a predefined. error message) Error_number : Error message: must be a user.

It’s value cannot be changed in the procedure body. parameter2 [mode] datatype. It is a named PL SQL block that can accept parameters(called arguments).Procedures A procedure is a type of sub-program that performs an action.. declaration section.. …] -. Cannot be assigned a default value. In OUT: Passed to the procedure. The Formal parameter acts as a constant and only takes the values passed by the calling environment. Actual parameters are the variables or expressions passed while calling the procedure Execute hikesal(v_empid. executable section and an optional exception-handling section. or an initialised variable. 2000). It can stored in the database as a schema object. Must be passed as a variable. End hikesal. It has a Header section. expression. Create or Replace Procedure procedure_name [parameter1 [mode] datatype. But it can take a default value. Cannot be assigned a default value. and then returned to the calling environment. The actual parameter can be a constant. . Formal versus Actual Parameters : Formal parameters are the variables declared in the parameter list of the procedure declaration Create or Replace procedure hikesal(p_empid number. OUT: Returned to the calling environment. Parameter Modes : IN(default) : Value passed to the procedure from the calling environment.header Is/AS --this section after is/as and before the begin is the declarative section PL SQl block --. Must be passed as a variable. p_empsal number) ….executable and an optional exception section The pl sql block starts with either a BEGIN or the declaration of local variables and ends with END or End procedure_name.

ve emp. end. end.put_line('Employee salary is '||vs). years number.empno. vsal from emp where empno=veno. vs). dbms_output. sal into vename.sal%type.put_line('Employee number is '||erec. vsal1 emp.hiredate%type. year IN number) is begin if year>=60 then vsal:=vsal*1. vsal OUT number) is begin select ename.Stored Procedure to display employee numbers and their salaries for department 10 create or replace procedure retemp(veno IN number. end if. end loop. dbms_output. vsal IN OUT number. ve. hiredate from emp. vdate1 emp.sal%type. end. . / to show increase in salary by 10 percent for all employees and an additional 5 percent for those who have completed more than 5 years of service create or replace procedure retemp(veno IN number.empno). / main block declare cursor c1 is select empno from emp where deptno=10.05.ename%type. sal.put_line('Employee name is '||ve). vename OUT varchar2. vs emp. begin for erec in c1 loop retemp(erec. dbms_output. / main plsql block declare cursor c2 is select empno.

/ at the calling environment : variable g_name varchar2(25) variable g_sal number variable g_comm. select months_between(sysdate. dbms_output. Number execute query_emp(100.begin for rec in c2 loop dbms_output. dbms_output. Methods of passing Actual Parameters : Positional : Here the actual parameters are passed in the order in which the Formal parameters are given with the procedure definition . p_comm. p_salary OUT employees.empno. p_name OUT employees.put_line('Employee salary '||vsal1).) print g_name print g_sal print The colon(:) is required to reference the host variable s in the execute command. years). p_salary. salary. dbms_output. :g_name.sal*1. commission_pct into p_name.vsal1.put_line('Employee salary '||rec. vsal1:=rec. p_comm from employees where employee_id=p_id. :g_sal.hiredate) INTO YEARS from emp where empno=rec. :g_comm.first_name%type.empno.salary%type.1.employee_id%type.commission_pct%type) is BEGIN Select first_name.empno). End emp_query. retemp(rec.put_line('Employee salary '||vsal1). end. end loop.sal). OUT employees.put_line('Employee number '||rec. / Viewing OUT Parameters : Create or replace procedure query_emp (p_id In employees.

. / Default values may be assigned only to In parameters . Default Option for parameters Create or Replace procedure add_dept(p_name IN departments. P_loc IN departments. End add_dept. ----positional add_dept(p_loc=>300. it must me the last item declared.location_id%type DEFAULT 100) IS Begin Insert into departments(department_id. Executing/Calling the above procedure : Begin add_dept. default End. Combination : Here the first actual parameter is passed positionally and the remaining are passed using the => symbol as in named association method.---Named association add_dept(p_loc=>400). end.Named Association : Here the actual parameters are passed in ant arbitary order by associating it with the corresponding formal parameter using the => symbol. and not to OUT or IN OUT Parameters. Because they are defined in the declaration section of another program/PL SQL Block/procedure. p_loc).put_line('P1 is called'). -.200). Any variable declared after the local subprogram will cause a compilation error. / In Combinational method all the positional parameters must be before the named parameters.department_name%type DEFAULT ‘unknown’.both default add_dept(‘Welfare’. or else there will be an error. location_id) values(departments_seq.nextval. declare procedure p1 is begin dbms_output. p_name=>’Medical’). department_name. their scope is limited to the parent(enclosing) block in which they are defined.will give an error as “a positional parameter may not follow a named parameter” Declaring Local Sub-procedures : The local procedures are not stored as Schema objects. Example : Execute add_dept(p_name=>’Legal’. p_name. 500) -. When you are declaring a sub-program in the declaration section of the main program.

End del_emp. procedure p1 is begin dbms_output. end. is never executed. p2. End log_emp. log)date) values(USER.put_line('p2 called'). dbms_output. / declare procedure p1. is not types as line 1) The result is an infinite loop.. / .put_line('In the main block').put_line('p1 called'). procedure p2 is begin dbms_output.begin p1. and dbms. Procedure to delete an employees record and to log the entry of the user who deleted the record Create or Replace procedure del_emp(p_id IN employees.employee_id%type) IS Procedure log_emp IS Begin Insert into log_table(user_id.. Begin Delete from employees where employee_id=p_id.put_line('In the main block').. end. log_exec. p2. / it will show error : p1 must be declared(if procedure p1. end. p1. end. begin p1. dbms_output. SYSDATE).

/ Handled Exceptions : When an exception is raised in a called procedure. the control goes immediately to it’s exception section. Begin For erec in emp_cursor Loop Raise_salary(erec.employee_id). Any DML statements issued before the exception was raised remain as part of the transaction. Commit. End loop. Commit. p_locid). End emp_proc. then the block terminates. ‘Mumbai’.Invoking a stored procedure from an anonymous block : Declare V_id number:=145. Consider the following tables LOCATIONS Location_ID City 1 Mumbai DEPARTMENTS Department_ID 5 Department_name Welfare Manager_ID 200 Location_ID 1 Create procedure Ins_dept(p_locid number) IS v_did number(4). 200. If the exception is handled. Begin Dbms_output. Begin Raise_salary(v_id). and the control goes to the calling program/environment. Select department_id into v_did from employees where employee_id=999.put_line(‘Procedure INS_dept started’). / . …… End. Insert into departments values(5. End. / Invoking a stored procedure from another Stored procedure : Create or replace procedure emp_proc IS Cursor emp_cursor is select employee_id from employees.

put_line(‘Invoking Procedure INS_dept’). the DML in ins_dept is not rolled back. which is not handled in procedure ins_dept. Procedure ins_dept selects the department id for an employee whose employee id=999. Dbms_output. the calling procedure terminates and the exception propagates to the calling environment. Ins_dept(p_lid). The procedure ins_dept is supoosed to insert a new department with department id=5. If the exception is not handled by the calling procedure. the control is passed to the calling procedure ins_loc where the exception no_data_found is handled.put_line(‘MAIN Procedure INS_LOC started’). So. then the DMLs in both the procedures are rolled back if no employee with employee_id=999 is found. Dbms_output. Manager ID=200 at the new location inserted through INS_loc. Suppose there is no such employee. Thus. ALL DML statements in the calling and called procedure are rolled back along with any changes to any host variables.put_line(‘No such employee’). all DML statements in the calling and called procedure remain as part of the transaction. / The procedure INS_loc inserts a new location into the locations table. v_dname varchar2(30). Select city into v_city from locations where location_id=p_lid. Begin Dbms_output. Insert into locations(location_id.put_line(‘Inserted City ’|| v_city). and becomes a part of the transaction of procedure ins_loc procedure Execute ins_loc(1. If in the above 2 procedures. ‘Mumbai’) If the exception is handled in the calling procedure. End. p_city). Exception When no_data_found then Dbms_output. p_city varchar2) IS v_city varchar2(30). city) values(p_lid. then it raises no_data_found. department name=Welfare.Create procedure Ins_loc(p_lid number. if the exception section from the ins_loc is removed. department/location for any .

put_line('Inside the function').put_line('value of x is'||x). If it is necessary to declare a variable it should be done in the declare part of the enclosing program declare x number. dbms_output. whereas a procedure returns zero or more values. parameter2 [mode] datatype. / Global or stored function : Stored in the database Defined and saved separately. x:=f1. executable part and an optional exception section. Like a procedure. Local functions : Not stored in database Defined and called within the same program Functions do not have declare. A function must return a value to the calling environment. a function must have a RETURN clause in the header and at least one RETURN statement in the executable section. end. end.header RETURN datatype Is/AS --this section after is/as and before the begin is the declarative section PL SQl block --.executable and an optional exception section End function_name. …] -. return 10. In addition. declarative section. Function is called as part of a SQL expression or as a part of a PL SQL expression. a function has a header. function f1 return number is begin dbms_output. can be called thru another pl/sql block . Create or Replace Function function_name [parameter1 [mode] datatype. begin dbms_output.put_line('Inside the Main').Functions Function is a named PL SQL block that can accept parameters and be invoked. / The type of parameter mode should be only of type IN.

End getsal. Begin Select salary into v_salary from employees where employee_id=p_id. to show the errors in the function coding.employee_id%type) RETURN number Is v_salary employee.Create or Replace Function getsal(p_id IN employees.put_line('salary is not found').salary%type :=0. . / at the sql prompt :>show_error. dbms_output.put_line('salary is '|| vsal). which is a good programming practice. as follows : variable gsalary number execute :gsalary :=getsal(130) print gsalary try the following on emp in scott : ed f1.sal%type. return true. create or replace function chkemp(veno number) return boolean is vsal emp. RETURN v_salary. exception when no_data_found then dbms_output. return false. to execute >@f1. / Executing functions : Declare a variable to hold the return value. begin select sal into vsal from emp where empno=veno. end.

sal from emp where sal<3000.main block declare y number :=&eno. inc number) return number is vsal number. select sal into vsal from emp where empno=veno.put_line('Employee Existing').empno.put_line('Employee Not Existing'). flag boolean. dbms_output. end. incr number :=&incr. else dbms_output. if flag=True then dbms_output.put_line('Increased salary is '||incsal||' for employee '|| erec. end. / . begin update emp set sal=sal+(sal*inc/100) where empno=veno. / stored function to update the salaries of employees getting less than 3000 with increment of certain % percent create or replace function upsal(veno number. incsal number. end if. return vsal. begin for erec in c1 loop incsal:=upsal(erec. end. / select text from user_source where type ='FUNCTION' and name='UPSAL' order by line main pl/sql block declare cursor c1 is select empno. incr).empno). begin flag:=chkemp(y). end loop. commit.

/ INVOKING a function from a SQL Expression Suppose you have a table employees with salary column then you may use the above function as follows to calculate their tax Select empno. salary. tax(salary) from employees where tax(salary) >(select max(tax(salary)) from employees where departmentid=30) order by tax(salary) desc Restrictions on calling function : The function must be a Stored Function It must accept only IN parameters Accept the parameters whose data type must be standard SQL data type and NOT a PL SQL data type . Advantages of Functions : To perform calculations that are complex. start with.Although 3 parameter modes IN. End Tax. order by and group by clauses Values clause of the insert command SET clause of the update command Select empno. Locations from where you may call a User-defined function : Select list of a SELECT command Condition of the WHERE and HAVING clause Connect by . do not use OUT and IN OUT with functions. Having a function returning multiple values is a poor programming practice. tax(salary) from employees. Increase efficiency of queries by performing functions in the query rather than in the application Create or Replace Function tax(p_value IN NUMBER) Return NUMBER IS Begin Return(p_value * 0. and it involves a lot of complications and side effects. since the purpose of a function is to accept 0 or more parameters and return a single value. awkward or unavailable with SQL Increase data independence by processing complex data at the Oracle server level(since the stored functions are stored at the server). OUT and IN OUT are available with functions. There can be a RETURN statement in the exception section also. rather than retrieving data into an application.08).

ALTER SYSTEM(system control statement). Examples Create or Replace function abc(p_sal number) Return number is Begin Insert into employees(employee_id. appoint_date. such a table is called MUTATING table . last_name.The return data types must be a standard SQL data type and NOT a PL SQL data type Restrictions from where a stored function can be called : Functions called from an Update/Delete statements on a Table XYZ cannot contain a DML statement on the table XYZ Functions called from an Update/Delete statements on a Table XYZ cannot query the table XYZ When called from an INSERT/UPDATE/DELETE/SELECT statement. the function cannot contain commands that end the transaction. End. salary) valurs(1001. sysdate-10. We are calling the function thru the update command on the same table employees. such a COMMIT(transaction control statement).’Pereira’. 10000). SET ROLE(session control statement). The function contains a DML statement on table employees. / Update employees set salary=xyz(100) where employee_id=95. If a Function internally calls sub subprogram(that is another function or procedure) then that sub program even muts not violate any of the above mentioned restrictions. Here also the mutating table error is returned When the code in a function is querying or reading the data in a table which is being updated. The update statement returns an error saying that the table is mutating Aslo Create or replace function xyz( a number) retutn number is s number. Begin Select salary into s from employees where employee_id=95. End. Return(p_sal +100). Return (s+a). or DDL commands Functions called from a SQL Expression(as shown above) cannot contain a DML Statement. / The function fails with the folliwng : Update employees set salary=abc(2000) where emplyee_id=90.

Procedure V/s.To drop a function : Drop function function_name. Function Executes as a PL SQL Block Invoked as part of an expressionm Exec Proc_name Does not contain a RETURN clause in Must contain a RETURN clause in the the Header Header May or may not return values Must return a single value Used to perform a series of actions Used to compute some value. which is returned to the calling environment .

DIRECT ACCESS From Hemant schema grant object –level privilge on the table employees to user Manoj Manoj has created the procedure manoj_proc that queries the Employees table in the schema Hemant INDIRECT ACCESS Manoj grants execute privilge to Saloni on his procedure manoj_proc This Saloni can access By default subprograms execute under security domain of the owner. drop.Managing Subprograms : System privileges are required to Create. but only through a procedure manoj__proc which is created by the user MANOJ. System privileges are granted by the user SYSTEM or SYS Grant Create (any) PROCEDURE to user_name. Grant Execute any PROCEDURE to user_name. The keyword ANY is optional for CREATE. Providing INDIRECT ACCESS to Data Suppose EMPLOYEES table in Scheme(user) HEMANT. not through a role If you are not the owner of a subprogram. alter. which is called DEFINER’s-RIGHTS. If a PL SQL subprogram refers to any objects that are not in the same schema. There is another user MANOJ and a third USER SALONI. Drop or Execute subprograms. The keyword ANY means that you ca create. Grant Drop any PROCEDURE to user_name. . Alter. then you must be granted access to those objects explicitly. or excute your own subprograms and those in another schema. Functions and Packages. Suppose we need to ensure that SALONI can access the table EMPLOYEES in the schema HEMANT. then you must have EXECUTE privilege on that subprogram. Grant Alter any PROCEDURE to user_name. The keyword PROCEDURE is used to assign such rights on Stored Procedures.

Package. USER_ERRORS Name -Name of the object Type -Type of object : Procedure. SHOW ERRORS without any [parameter will show the compilation errors The following query will also show you an output which is similar to that of show errors . Package. PACKAGE BODY. ALL_SOURCE and DBA_SOURCE contain the additional column Owner Select text from user_source where name=’PROCEDURE_name’ order by line. TRIGGER -Date of object creation -Date when object was last modified -Date and time when object was last recompiled -VALID or INVALID You can also examine the ALL_OBJECTS and DBA_OBJECTS. query the user_objects USER_OBJECTS Object_name Object_Id Object_type CreateLast_DDL_Time Timestamp Status -Name of Object -Internal identifier for the object -Whether TABLE.User_Objects To obtain names of all PL/SQL stored objects within a schema. PACKAGE. Package Body LINE -Line number of the source code where error occurred Position -Position in the line where the error occurred Text -Text of the error message SHOW ERRORS [Function/Procedure/Package/Package body/Trigger/View This will show use the errors that occurred while trying to compile the articular program unit. Function. Package Body LINE -Line number of the source code TEXT -Text of the source code line This table contains one record for each line of code. PROCEDURE. both of which contain an additional OWNER column. USER_SOURCE Name -Name of the object Type -Type of object : Procedure. Function. FUNCTION.

Select line || ‘/’||position. . text from user_errors where name=’proc_name’ order by line.

TRIGGERS Definition : Trigger is basically a PL/SQL block or PL/SQL procedure associated with a Table/View/Schema/Database and which executes implicitly whenever a particular event takes place Types of Triggers : Application Triggers which execute implicitly when a particular DML event occurs within an application such as Oracle Forms Developer Database Triggers which could be 1) Data Event or 2) System Event Data Event trigger fires implicitly whenever a DML operation occurs on table or a DDL event such as CREATE/ALTER are issued. It could also be an INSTEAD OF Trigger on a View. They are fired irrespective of the user connected and irrespective of the application used. System Event Triggers are fired when a system event occurs such as user logs on, user shutsdown, etc Guidelines for Designing Triggers Dos Design Triggers to guarantee that when a specific operation is performed, related actions are performed. Use Triggers only for Centralised, Global Operations that occur for the triggering event, regardless of the user or application that issues the triggering statement. Donts Do not use triggers when the functionality is already built into Oracle Server. Do not use triggers to implement integrity constraints. Use the built-in constraints Excessive use of triggers can result into complex interdependencies. If the logic is very lengthy, write a stored procedure for the same and invoke the stored procedure in the trigger body If multiple triggers of the same type are defined for a table, their order of execution is arbitary. To ensure that multiple triggers of the same type are executed in a particular order, write stored procedures for the different actions and have a single consolidated trigger in which you call those stored procedures in the required order. The Trigger Statement contains : Trigger Timing :BEFORE, AFTER for tables Triggering Event :INSERT, UPDATE or DELETE Table name :On table or view Trigger Type :Statement level or Row level Trigger Body :PL SQL block INSTEAD OF for views

BEFORE :Execute the trigger body before the triggering DML event on the table To determine whether the triggering statement should be allowed to complete To calculate some values before completing the triggering Insert or Update statement To validate or set some complex business logic To initialize some global variables or flags BEFORE :Execute the trigger body before the triggering DML event on the table To complete the triggering statement before executing the trigger body INSTEAD OF This type of trigger is used to modify data through a view that was otherwise not modifiable because of the inherently non-modifiable nature of the view. Such trigger works in the background by performing the action directly on the underlying base tables involved in the view. You may write INSERT, UPDATE or DELETE statements on a view. The Triggering Event can be INSERT, UPDATE or DELETE statement on a table. In case of UPDATE statement, you can specify the column, which if changed, causes the trigger to fire. The triggering event can contain one or two or all the three DML operations. The trigger type can be Statement Trigger or Row Trigger. A Statement Trigger is fired once for the triggering event. It may affect zero or more rows, but is fired only once. They are used when the trigger-action to be performed is independent of the data affected by the triggering event. A Row Trigger is fired once for each affected row. If no rows are effected, then the trigger is not fired. They are used when the trigger-action depends on the data or rows affected by the triggering event. The Trigger body defines the action that needs to be performed when the triggering action is issued. It is a PL SQL block containing SQL and PL SQL statements defining PL SQL types, item, variables, constants, cursors, exceptions, etc. You can call other PL SQL procedures or Java Procedures. Trigger size cannot be more than 32K. The Firing Sequence is as follows : BEFORE Statement Trigger -------- once BEFORE Row Trigger AFTER Row Trigger AFTER Statement Trigger -------- once for each affected row -------- once for each affected row --------- once

Syntax for Statement Level Trigger Create or Replace Trigger Trigger_name Timing Event1 or Event2 or Event3 On table_view_name Trigger_body The trigger name must be unique with respect to other triggers in the same schema. Trigger name need not be unique with respect to other schema objects like table, view, or procedure. A trigger to prevent insertion of records into the dept table on Saturdays or Sundays or on non-working hours on other days create or replace trigger check_deptno before insert on dept begin if (to_char(sysdate,'DY') in ('SAT','SUN')) or (to_char(sysdate,'HH24:MI') not between '09:00' and '17:00' ) then raise_application_error(-20500,'Insert Not allowed'); end if; end; / Raise_application_error is a server-side built-in procedure that returns an error to the user and causes the PL SQL block to fail. When the database trigger fails, the triggering statement is automatically rolledback by Oracle Server. Table : DEPT DEPTNO varchar2 DEPTNAME vrachar2 A trigger to prevent insert or update or delete operations on table dept create or replace trigger check_deptno before insert or update or delete on dept begin if deleting then raise_application_error(-20500,'Delete Not allowed'); elsif Inserting then raise_application_error(-20501,'Insert Not allowed'); elsif Updating('deptno') then raise_application_error(-20502,'Updating of deptno Not allowed'); else raise_application_error(-20503,'Updating of Table Not allowed'); end if; end; /

'Hello'). If the condition is satisfied then only we enter into the trigger body. Old_deptno varchar2(15). / with the WHEN clause you can specify a condition in the brackets.deptno>50) begin raise_application_error(-20555. The condition is evaluated for each row to determine whether or not the trigger body is executed. Time timestamp. Thus. The WHEN clause is optional. Operation For INSERT For UPDATE For DELETE Old Value NULL Value before update Value before update New Value Newly Inserted value Value after update NULL Table : AUDT_DEPT create table audt_dept ( User_name varchar2(15). UPDATING. DELETING. Old_deptname varchar2(15). UPDATING(‘column’) can be used in a single trigger for combining several triggering events into one single trigger Syntax for Row Level Trigger Create or Replace Trigger Trigger_name Timing Event1 or Event2 or Event3 On table_view_name For Each Row [When (condition)] Trigger_body create or replace trigger check_dept before insert on dept for each row when(new. we can use the OLD and NEW Qualifiers to reference the value of a column before and after the data change by prefixing the column name with the OLD and NEW Qualifiers. You need to use the colon(:) as a prefix in every SQL and PL SQL statement. the WHEN clause is to restrict the Trigger action to those rows that satisfy a certain condition With ROW Level trigger only.Special Conditional Predicates INSERTING. end. New_deptname varchar2(15) ) . New_dept varchar2(15). No colon(:) required if the Qualifiers are used in the WHEN clause.

deptno.deptname). The INSTEAD OF trigger must enforce the check.deptname. Such triggers are Row Level Triggers. group by clause. end. Create or Replace Trigger Trigger_name INSTEAD OF Event 1 or event2 or event3 On View_name For each row Trigger_body Even if FOR EACH ROW is omitted. distinct operator or a join of 2 or more tables. INSTEAD OF Trigger It is used to modify the data where a DML statement has been issued against an inherently non-updatable view(that is views which are non-updatable because the underlying query may contain set operator. :old. :new. INSTEAD OF Triggers is still defined as a Row Trigger. sysdate.create or replace trigger audt_dept after insert or update or delete on dept for each row begin insert into audt_dept values(user. this trigger is fired by Oracle sever instead of executing the triggering statement. :new. start. Unlike other triggers. Consider the following tables : Table Empid Empname Salary Deptno Table Deptno Deptname Totalsal : Employee : Dept_stat . BEFORE and AFTER options are not valid. :old. The WITH CHECK option for views is not enforced when insertion/updation are performed by using the INSTEAD OF Trigger. group function.deptno. connect by.

where Elsif UPDATING(‘SALARY’) then Update employee set salary=:new.deptno.salary where empid=:old.deptno).Following trigger updates the total salary in Dept_stat table whenever rows are inserted/update/deleted from the Employee table.empname. e.empid.deptno.deptno. :new.salary .empname. Update dept_stat set totalsal=totalsal + :new.salary. Create or Replace Trigger New_emp_check INSTEAD OF INSERT OR DELETE OR UPDATE FOR EACH ROW Begin If INSERTING THEN Insert into employee values(:new.deptno where empid=:old.deptno.:old. / To Enable/Disable a Trigger Alter Trigger Trigger_name DISABLE/ENABLE To Enable/Disable all Triggers on a table Alter Table Table_name DISABLE/ENABLE ALL TRIGGERS To Recomplie a Trigger Alter Trigger Trigger_name COMPILE . :new. dept_stat d where e. Update dept_stat set totalsal=totalsal-:old.salary where deptno=:old.salary where deptno=:new.deptno.empid. End if. Update dept_stat set totalsal=totalsal +(:new. :new. End.salary deptno=:old.deptno. Update dept_stat set totalsal=totalsal :old.deptno=d.salary where deptno=:old.salary where deptno=:new.empid. Create view empl_view as Select e.salary. Update dept_stat set totalsal=totalsal+:new.empid. e.deptno. e.empid.deptname from employee e. d. Elsif Updating(‘Deptno’) then Update employee set deptno=:new. Elsif DELETING Then Delete from employee where empid=:old.

Events that can cause a System Level Trigger to fire are : A DDL statement on a object in the database or schema (the trigger can be Database Level or Schema Level) Specific User or any other User logs on or off (the trigger can be Database Level or Schema Level) Database Shutdown or Startup (the trigger can be Database Level only Triggers on DDL Statements : Create OR Replace Trigger Trigger_name Timing DDL_Event1 Or DDL_Event2 or…. Synonym. Table. Procedure.This command is used to explicitly recompile a trigger. Function. regardless of whether it is valid or invalid. On Database/Schema Trigger_body Timing : BEFORE or AFTER DDL_Event : CREATE OR ALTER OR DROP. Trigger. Tablespace. Index. The DDL Triggers fire when the Object is a Cluster. Type . Sequence. Triggers A Database Level Trigger fires for all users. A Schema or Table Level Trigger fires only when that schema or table is involved. On Database/Schema : You can create triggers on DDL Statements at the DATABASE or SCHEMA Level. To Drop a Trigger Drop Trigger Trigger_name. Package. On Database/Schema Trigger_body . All the triggers on a table are dropped when the table is dropped. SYSTEM LEVEL TRIGGERS Triggers on system events can be defined at the Database Level or Schema Level. View or a User Triggers on System Events : Create OR Replace Trigger Trigger_name Timing Database_Event1 Or Database_Event2 or…. Trigger_body : It is a complete PL SQL Block. Role. Causes the Oracle Server to fire the trigger whenever a Create/Alter/Drop Statement modifies a Database Object in the Data Dictionary.

salary) MUTATING TABLE When the code in a trigger is querying or reading the data in a table which is being updated. / CALL Statement : It is recommended to create stored procedures and call them in the Trigger body rather than coding the PL SQL body in the trigger itself. The CALL statement enables you to call a stored procedure. There is no semicolon at the end of the CALL Statement. ‘Logged On’). END. but SHUTDOWN and STARTUP apply only to the Database Level. :new. AFTER LOGON : Whenever a User Logs on to the Database BEFORE LOGOFF : Whenever a User Logs off the Database AFTER STARTUP : Whenever a Database is opened BEFORE SHUTDOWN: Whenever a Database is shut down These triggers can be created at the DATABASE or SCHEMA LEVEL. END. such a table is called MUTATING table.Database_Events : AFTER SERVERERROR : Whenever a Server Error Message is logged. Create or Replace Trigger salary_check Before Update of Salary . C. The procedures so called can be implemented in PL SQL. job_id ON Employees For Each Row When(new. Java. date. Reading and Writing data using triggers follows certain rules.job_id. ‘Logged Off’). SYSDATE. INSERT OR DELETE statement and a ROW .job_id<>’YYZ’) CALL check_sal(:new. / CREATE OR REPLACE TRIGGER logoff_trig AFTER LOGOFF ON SCHEMA BEGIN Insert into log_trig_table(userid. action) values(USER. date. SYSDATE. A mutating table is the one that is currently being modified by an UPDATE. action) values(USER. CREATE OR REPLACE TRIGGER logon_trig AFTER LOGON ON SCHEMA BEGIN Insert into log_trig_table(userid. which are applicable to only row level triggers or a statement level trigger that is fired because of the referential integrity constraint ON DELETE CASCADE.

referencing_names.salary<v_minsal or :New.salary>v_maxsal then Raise_application_error(-20505. max(salary) into v_minsal. User_errors : Contains details of the compilation errors that occurred while the trigger was compiling . V_maxsal employees. BEGIN Select Min(salary). Consider a trigger which checks that the salary for a new employee or an existing employee is always between the minimum and maximum salaries for that job code. Create or Replace Trigger Check_Sal Before INSERT Or UPDATE OF salary. trigger_type. trigger body. created from user_objects where object_type='TRIGGER' User_Triggers : Contains name.job_id IF :New. table on which table was created. status. type. triggering_event.level trigger on any such event is trying to read/write data into the same table. select object_name.’Salary Out Of Range’). and any other table referencing it with the FOREIGN KEY constraint is also mutating. (A table is not mutating for STATEMENT level triggers. trigger_body from user_triggers table_name. / Viewing Trigger Information User_objects : Contains name and status of the trigger and the date and time when the trigger was created.) A triggered table itself is mutating. triggering event. status. v_maxsal from employees where job_id=:new.salary%type. select trigger_name. END IF: END.salary%type. job_id On Employees When (new.job_id <>’PROG’) Declare V_minsal employees. subobject_name.

which are stored separately in the database. exceptions or types. Variables declared in the package specification are initialized to NULL by default. The Replace option drops and recreates the package specification. The public procedures or functions can be invoked repeatedly by other constructs in the same package or from outside the package. generally. has two parts : Package Specification and Package Body. Package Specification is the INTERFACE to your application. Public type and item declarations : Declares variables. The package cannot be itself parameterized. Can only be referenced by other constructs which are part of the same package. and thus implements the specification.PACKAGES It is used to group together related PL SQL Types. Creating the Package Specification Create or Replace Package package_name Is/as Public type and item declarations Subprogram specifications End package_name. cursors and sub programs available for use Package Body fully defines the cursors and subprograms. Private or Local : Declared and Defined within the package body. types. cursors. Subprogram specification : Declares the PL SQL subprograms. Further calls to other constructs of the same package do not need any disk I/O. called or nested. items. constants. Can be referenced and changed(in case of variables) outside the package and are visible to the external users. A Package. In the Specification we declare the variables. the entire package gets loaded into memory. This allows to change the definition of a program construct in the package body without causing the Oracle server to invalidate the other schema objects that call or reference the program construct. exceptions. subprograms(procedures and functions) into one single container. Types of Package Constructs : Public or Global : Declared within the package specification and may be defined in the package body. When you call a packaged construct for the first time. . constants.

10. The function should be called through a procedure in the same package. Create or Replace Package comm_package Is g_comm Number :=0. If p_comm>v_max_comm then return(FALSE). / Create or Replace Package Body comm_package Is Function validate_comm(p_comm IN Number) Return Boolean Is v_max_comm Number. public and private. constants. if a package specification does not declare subprograms and cursors and declares only types. Begin Select max(commission_pct) into v_max_comm from employees. Thus. then the package body is not necessary. END comm_package. Subprograms and Cursors are declared(without their respective bodies) in the package specification. . and they have their underlying implementation(or bodies) in the package body. constants. All private constructs must be declared before they are used in the public constructs. The Replace option drops and recreates the package body. End if. variables. Subprogram Bodies : Defines the PL SQL subprograms. But the package body can still be used to initialize items declared in the package specification.declared and initialized procedure reset_comm(p_comm IN Number). cursors. End validate_comm. Else return(TRUE). exceptions or types. Private type and item declarations : Declares variables. Write a package that contains a function that validates the commission so that the commission may not be greater than the highest commission amongst all the employees. and this procedure should display the appropriate messages and reset and validate the prevailing commission. exceptions.Creating the Package Body Create or Replace Package Body package_name Is/As Private type and item declarations Subprogram Bodies(Private and Public) End package_name. -.

/ When you invoke a package or function from within the same package. End if. This package specification declares public(global) variables that exist for the duration of the user session.Procedure reset_comm(p_comm IN NUMBER) Is Begin If validate_comm(p_comm) then g_comm:=p_comm.’Invalid Commission’).15) To invoke a procedure from a package through a different schema Execute scott.reset_comm(0.15) To invoke a procedure from a package through a remote database Execute comm_package.comm_package.put_line(’20 miles = ‘ ||20*global_consts.25) To invoke a procedure from a package through iSQL *Plus Execute comm_package. / Execute dbms_output.reset_comm(0.reset_comm@ny(0.936.reset_comm(0. you need not qualify it’s name with the package name. End comm_package.mile_2_kilo|| ‘ kms.9144.15) Declaring Bodiless Package Create Or Replace Package global_consts IS mile_2_kilo constant number :=1. The above function is a private function(declared and defined within the package body). To invoke a package procedure or function from outside the package. you need to qualify it’s name with the name of the package Comm_package. meter_2_yard constant number :=1. Else Raise_application_error(-20210. yard_2_meter constant number :=0. kilo_2_mile constant number :=0.’).6214.6093. . end global_consts. End reset_comm. hence it can be referenced by only other procedures/functions of the same package.

Overloading is a feature which enables you to have different packaged subprograms with the same name. since multiple subprograms of the same name are allowed Better Performance : Entire package loaded into memory when a package is first referenced Only one copy of the package in memory for all the users Dependency hierarchy is simplified Overloading support in Packages Packaged subprograms(procedures or functions) Overloading not possible for stand-alone subprograms.:yard) > print yard To Drop the package Specification : Drop Package package_name. Added Functionality since it allows persistency of variables and cursors Overloading allowed.Referencing a Public variable from a stand-alone procedure Create or Replace procedure meter_to_yard(p_meter IN number. p_yard Out number) Is Begin p_yard:=p_meter * global_consts. To Drop the package body : Drop Package body package_name. .meter_2_yard. Hiding Information : Only package declarations made in the specification are visible and accessible to applications Private package constructs are hidden and inaccessible All the coding is hidden in the package body. End meter_to_yard. Advantages : Modularity since it encapsulates related constructs Easier Application Design. / at the sql prompt >variable yard number >execute meter_to_yard(1. since specification and body are coded and compiled separately. order or data-type of the formal parameters of the subprograms. can be overloaded. but differing in the number.

empage) values(p_empno. order and datatypes of the actual parameters passed with the called subprogram with formal parameters of the one or more subprograms searched in a scope. the compiler stops the search.nextval. End add_employee. p_empname. Resolving a Subprogram call When a call is made to a subprogram. . p_empage). Example : NUMBER and DECIMAL datatypes belong to the same family Subprograms which differ only in the subtypes of their parameters. Procedure add_employee(p_empname IN employees. Example : VARCHAR and STRING are PL/SQL subtypes of VARCHAR2. END overload. empname. p_age IN employees.empname%type default ‘unknown’.empname%type default ‘unknown’.empage%type default 21).RESTRICTIONS 1) 2) 3) Standalone subprograms cannot be overloaded Subprograms which differ only in the datatypes of their parameters. but the different subtypes belong to the same family. p_empage).empno%type. p_age IN employees. empname. As soon as the compiler finds one or more subprograms with the required name.empno%type. p_empname IN employees.empname%type default ‘unknown’. it starts searching in the enclosing scopes. Procedure add_employee(p_empname IN employees. p_age IN employees. p_empname. It then starts matching the number. empage) values(empnoseq.empage%type default 21) Is Begin Insert into employees(empno. p_empname IN employees. p_age IN employees. but the different datatypes belong to the same family. the compiler starts searching for a subprogram with that name in the current scope. Create or replace package body overload Is Procedure add_employee(p_empno IN employees.empage%type default 21) Is Begin Insert into employees(empno.empage%type default 21).empname%type default ‘unknown’. Create or replace package overload Is Procedure add_employee(p_empno IN employees. and if it does not find one in the current scope.

then to use the built-in to_char function . whereby you give the subprogram specification terminated by a semicolon.to_char Using Forward Declarations PL/SQL does not allow forward referencing. If you redeclare a built-in subprogram as a stand-alone subprogram then to access your sub-program you need to qualify it with your schema.add_employee(101. since public package constructs are already declared in the package specification. Thus. 23) If you redeclare a built-in subprogram in another PL SQL Program unit. To solve this. example scott. you call it as :STANDARD. since the subprogram specifications go in the package specification.TO_CHAR.End add_employee END overload. To access the built-in subprogram.’Rahul’. At the SQL PROMPT >Execute overload. Thus packages support forward declarations. if you have redeclared the to_char function.add_employee(’Rahul’. because the function to_char belongs to the builtin package STANDARD. your local declaration overrides the built-in or standard subprogram. you need to qualify it with the package name. The subprogram can be then defined after another subprogram that calls it. . which means that a subprogram cannot be referenced or called before it has been declared. and the subprogram bodies go in the package body. Forward declarations are used when you need to : Define subprograms in a logical or alphabetical order Define mutually recursive subprograms(which call each other directly/indirectly) Group subprograms in a package Use Forward Declarations in the cases of private constructs or private subprograms of the package. PL/SQL allows a special subprogram declaration called Forward declaration. The formal parameter list must appear in both the forward declaration and the subprogram body. where they are invisible to the applications. 32) >Execute overload. but both must appear in the same program unit. The subprogram body can appear anywhere after the forward declaration .

--forward declaration Procedure award_bonus(…. In such cases. / .) Is begin Calc_rating(…) End. ----declare all other public procedures/functions/variables END otop.No End for this one-time-only procedure END otop. Procedure calc_rating(…) Is Begin End.). we use one-time-only procedures. / Create or Replace Package Body otop Is ----declare all private procedures/functions/variables ----define private/public procedures/functions Begin Select rate into hra from rate_mast where rate_name=’HRA’. End forwardpack..Create or Replace Package Body forwardpack Is Procedure calc_rating(. ----. Create or Replace Package otop Is Hra number. because the value is reset by the one-time-only procedure. since such procedures are executed only once when the package is invoked within a user session. The keyword END is not used at the end of the One-Time-Only procedures. For such type of variables do not initialize them in the declaration. / One-time-Only Procedure When the derivation of the values is too complex it is difficult to initialize public or private variables to that value in the variable declaration.

End if.tax_calc(salary).Restrictions on Package Functions 1) A function called from a query or DML statement cannot end the current transaction. create or rollback to a savepoint or alter the system session 2) A function called from a query statement or a DML statement cannot execute a DML statement or otherwise modify the database 3) A function called from a DML statement cannot read or modify the table on which the DML was fired. End validate_comm. / To use the packaged function : Select tax_calc_pack. End tax_calc_pack. PERSISTENT State of a Package Create or Replace Package comm_package Is g_comm number :=10. End tax_calc. Create or Replace Package tax_calc_pack Is Function tax_calc(p_value in number) return number. end comm_package. Else return(FALSE). Begin Select max(commission_pct) into v_max_comm from employees. Begin return(p_value*v_rate). procedure reset_comm(p_comm in number). If v_max_comm >=p_comm then return(TRUE). / Create or Replace Package Body tax_calc_pack Is Function tax_calc(p_value in number) return number Is v_rate number :=0. . End tax_calc_pack. / Create or Replace Package Body comm_package Is Function validate_comm(p_comm in number) return Boolean Is v_max_comm number.08. salary from emp.

5) Now maximum is 0. Max_comm=0.3 1000 1100 1145 . / Time 0900 User X Execute comm_package.reset_comm(0.reset_comm(0. Exit Logged In again Since the earlier transaction was rolled back. Invalid Commission Rollback .5 is Invalid Hence.4. Max_comm=0.25 Hence g_comm=0. Max_comm=0.4>0.4>0.Procedure reset_comm(p_comm IN number) Is Begin If validate_comm(p_comm) then g_comm:=p_comm.0.6) For this user.9.9>0.’Invalid Commission’).3 Hence g_comm=0. End if. since user Y did not complete his transaction. maximum is still 0.5 Hence g_comm=0.reset_comm(0. End comm_package.4.25 User Y 0930 0935 Insert into employees(empno.4 Execute comm_package. max_comm is 0.9) Execute comm_package.5 Execute comm_package.4>0. End reset_comm.25) Assume maximum is 0.reset_comm(0. commission_pct) values(‘Ram’.3) Max_comm=0. Else raise_application_error(-20210.

Exit when c1%rowcount>=3. End loop. End pack_cur. The state does not persist across several sessions of the same user or acroos different users. / The cursor is opened in P1. the next three rows are fetched and displayed. Procedure p2. Loop Fetch c1 into v_empno. Close c1. / Create or Replace Package Body pack_cur Is v_empno number.Persistent State of a Packaged Cursor Create or Replace Package pack_cur Is Cursor c1 is select empno from emp order by empno desc. When we continue to fetch in P2. End pack_cur.put_line(‘Employee Number :’ || (v_empno)). End p2. The persistent state of a cursor/variable is maintained in a session of a user. . Procedure p2 is Begin Loop Fetch c1 into v_empno. Procedure p1. Thus packages maintain a persistent state of the package cursors. Dbms_output. Exit when c1%rowcount>=6. Procedure p1 is Begin Open c1. End p1. End loop. Dbms_output.put_line(‘Employee Number :’ || (v_empno)). three rows are fetched and displayed. and the cursor is not closed.

NCLOB.LOB LOB is a data type to store large. unstructured data such as text. based on the byte-length of the national character set To interact with LOB. A BLOB< CLOB or NCLOB can be used in one of the following : Column of a table Attribute of a user-defined type PL SQL variable. photo BLOB). etc. Internal LOBs : They are stored inside the Oracle server in the LOB segment. fname varchar2(25). video clippings. but the LOB locator is stored in the row. . CLOB is interpreted by the Oracle server as a single-byte character stream NCLOB is interpreted by the Oracle server as a multiple-byte character stream. There are 4 large object data types BLOB for Binary large objects such as video clip CLOB for character large objects NCLOB for multibyte chracter large objects BFILE for a binary file stored at the OS level outside the database. BFILEs can be accessed only in the read-only mode from an Oracle server Components of a LOB : The 2 distinct components of an LOB are : LOB value : It is the data that constitutes the real object being stored LOB locator : It is a pointer to the location of the LOB value stored inside the database The LOB value may be stored anywhere in the database. similar to LONG RAW. parameter or return value(result) BLOB is interpreted by the Oracle server as a bit-stream. The BFILE column stores a file locator that points to the external file LOBs are characterized in two ways : Internal LOBs(CLOB. BLOB) stored in the database External files(BFILE) stored outside the database. graphic images. resume CLOB. the following interfaces are used : PL/SQL package DBMS_LOB Oracle Call Interface(OCI) Oracle Objects for Object Linking & Embediing(OLE) Pro*C/C++/Pro*COBOL precompilers JDBC Suppose you have the following table : Create table employee (emp_id number(5). lname varchar2(25).

Before you start writing data to CLOB/BLOB columns using DBMS_LOB package or OCI(oracle call interface) .D’. -. Updating CLOB using DBMS_LOB in PL/SQL : Declare lobloc clob. MCSD. MCSD.’ ‘Manoj’. MCP. offset. -. values(10.’ Pangam’. OCP’.write at lob location. -. If the record is already there. ‘M. fname varchar2(25). it must contain a locator that points to an empty or populated LOB value. amount. empty_clob().write(lobloc. photo BLOB default empty_blob()).Tech. Insert into employee values(20. amount to be written. NULL).where to start writing Begin Select resume into lobloc from employee where emp_id=10 FOR Update.MCDBA. text). --serves as the LOB locator text varchar2(30000):='Corporate Trainer'. amount :=length(text). ‘Pangam’. You may initialize CLOB and BLOB columns by using the functions EMPTY_CLOB() and EMPTY_CLOB() functions respectively. that is.getlength(lobloc)+2. ‘Hemant’.MCDBA. ‘Hemant’. Pangam’. This can be done through the insert or update statements as follows : Insert into employee empty_blob()). OCP Ph. NULL). Now we can put values as follows : Insert into employee values(10. resume CLOB default empty_clob(). Now the columns have been initialized. dbms_lob. then you could have used UPDATE command. such columns must be made non-null. ---For update locks the row and returns the LOB locator for the ---resume lob column offset :=DBMS_LOB. MCP. but not yet populated by data. You could have done the initialization work at the time of table definition as follows : Create table employee (emp_id number(5). ‘MCA. Select resume from employee. amount number.amount of data to be written offset number. lname varchar2(25). from where and text to write .

text). Col2 BLOB). To disassociate a LOB value from a row: Update employee set resume =empty_clob() where emp_id=40 To Add LOB columns Alter table table_name add(col1 CLOB.instr(resume. / Selecting CLOB values using DBMS_LOB select dbms_lob. -. dbms_lob.bmp')) . dbms_lob. create the following table : create table x(col1 varchar2(10).'c1. amount. bfilename('saloni'. End. Managing BFILES Create a directory ‘saloni’ as a oracle user and give the path of the directory which contains your files as follows : Create or replace directory SALONI as ‘d:\saloni\saloni’ Give read privileges to users on that directory as follows : grant read on directory SALONI to public. End.writeappend(lobloc. 5).bmp and c2.substr(resume.put_line(‘Resume is : ‘ || text).bmp')).'c2. insert into x values(‘c2’. Use the BFILENAME function to inset values as follows. 'Cor') from employee Selecting CLOB values in PL/SQL Declare Text varchar2(4000). col2 bfile).bmp are existing in the folder : insert into x values(‘c1’. Dbms_output. Amount:=length(text) . Begin Select resume into text from employee where emp_id=10.text:='Corporate Consultant cum System Analyst'. 5. bfilename('saloni'. amount to be written and text to write commit. / Removing LOBs To delete row containing LOBs : Delete from employee where emp_id=30. ensuring tha the files c1. Select resume into lobloc from employee where emp_id=20 FOR Update.append at lob location.

update x set col2=v_file where current of emp_cursor. Begin For erec IN emp_cursor Loop v_filename:=erec. cursor emp_cursor is select col1 from x for update.bmp'.put_line('Loaded File :' || v_filename ||' size : '|| dbms_lob. v_file:=BFILENAME(p_file_loc. dbms_lob. End loop. v_filename). / . dbms_lob. you may now use subsequent read operations on BFILE using the PL/SQL dbms_lob package and OCI.fileopen(v_file).getlength(v_file)).fileclose(v_file).Once physical files are associated with the records using SQL.col1||'. v_filename varchar2(20). Then the following procedure is used to load a BFILE pointer to an image of each employee into the table ‘X’ using DBMS_LOB package Create or Replace procedure load_emp (p_file_loc IN varchar2) IS v_file BFILE. dbms_output. End load_emp.