Exception Handling

In this section we will discuss about the following,

1) What is Exception Handling.

2) Structure of Exception Handling.

3) Types of Exception Handling.

1) What is Exception Handling?

PL/SQL provides a feature to handle the Exceptions which occur in a PL/SQL Block known as exception Handling.
Using Exception Handling we can test the code and avoid it from exiting abruptly. When an exception occurs a messages which explains its cause is recieved.
PL/SQL Exception message consists of three parts.
1) Type of Exception
2) An Error Code
3) A message

By Handling the exceptions we can ensure a PL/SQL block does not exit abruptly.

2) Structure of Exception Handling.

The General Syntax for coding the exception section

 DECLARE

   Declaration section 

 BEGIN 

   Exception section 

 EXCEPTION 

 WHEN ex_name1 THEN 

    -Error handling statements 

 WHEN ex_name2 THEN 

    -Error handling statements 

 WHEN Others THEN 

   -Error handling statements 

END; 

General PL/SQL statments can be used in the Exception Block.

When an exception is raised, Oracle searches for an appropriate exception handler in the exception section. For example in the above example, if the error raised is 'ex_name1 ', then the error is handled according to the statements under it. Since, it is not possible to determine all the possible runtime errors during testing fo the code, the 'WHEN Others' exception is used to manage the exceptions that are not explicitly handled. Only one exception can be raised in a Block and the control does not return to the Execution Section after the error is handled.

If there are nested PL/SQL blocks like this.

 DELCARE

   Declaration section 

 BEGIN

    DECLARE

      Declaration section 

    BEGIN 

      Execution section 

    EXCEPTION 

      Exception section 

    END; 

 EXCEPTION

   Exception section 

 END; 

In the above case, if the exception is raised in the inner block it should be handled in the exception block of the inner PL/SQL block else the control moves to the Exception block of the next upper PL/SQL Block. If none of the blocks handle the exception the program ends abruptly with an error.

3) Types of Exception.

There are 3 types of Exceptions.


a) Named System Exceptions

b) Unnamed System Exceptions

c) User-defined Exceptions

a) Named System Exceptions

System exceptions are automatically raised by Oracle, when a program violates a RDBMS rule. There are some system exceptions which are raised frequently, so they are pre-defined and given a name in Oracle which are known as Named System Exceptions.

For example: NO_DATA_FOUND and ZERO_DIVIDE are called Named System exceptions.

Named system exceptions are:

1) Not Declared explicitly,

2) Raised implicitly when a predefined Oracle error occurs,

3) caught by referencing the standard name within an exception-handling routine.

For Example: Suppose a NO_DATA_FOUND exception is raised in a proc, we can write a code to handle the exception as given below.

BEGIN 

  Execution section

EXCEPTION 

WHEN NO_DATA_FOUND THEN 

 dbms_output.put_line ('A SELECT...INTO did not return any row.'); 

 END; 

b) Unnamed System Exceptions

Those system exception for which oracle does not provide a name is known as unamed system exception.
These exception do not occur frequently. These Exceptions have a code and an associated message.

There are two ways to handle unnamed sysyem exceptions:


1. By using the WHEN OTHERS exception handler, or

2. By associating the exception code to a name and using it as a named exception.

We can assign a name to unnamed system exceptions using a Pragma called EXCEPTION_INIT.


EXCEPTION_INIT will associate a predefined Oracle error number to a programmer_defined exception name.

Steps to be followed to use unnamed system exceptions are


• They are raised implicitly.

• If they are not handled in WHEN Others they must be handled explicity.

• To handle the exception explicity, they must be declared using Pragma EXCEPTION_INIT as given above and handled referecing the user-defined exception name in the exception section.

The general syntax to declare unnamed system exception using EXCEPTION_INIT is:

DECLARE 

   exception_name EXCEPTION; 

   PRAGMA 

   EXCEPTION_INIT (exception_name, Err_code); 

BEGIN 

Execution section

EXCEPTION

  WHEN exception_name THEN

     handle the exception

END;

For Example: Lets consider the product table and order_items table from sql joins.

Here product_id is a primary key in product table and a foreign key in order_items table.

If we try to delete a product_id from the product table when it has child records in order_id table an exception will be thrown with oracle code number -2292.

We can provide a name to this exception and handle it in the exception section as given below.

 DECLARE 

  Child_rec_exception EXCEPTION; 

  PRAGMA 

   EXCEPTION_INIT (Child_rec_exception, -2292); 

BEGIN 

  Delete FROM product where product_id= 104; 

EXCEPTION 

   WHEN Child_rec_exception 

   THEN Dbms_output.put_line('Child records are present for this product_id.'); 

END; 

/ 

c) User-defined Exceptions

Apart from sytem exceptions we can explicity define exceptions based on business rules. These are known as user-defined exceptions.

Steps to be followed to use user-defined exceptions:


• They should be explicitly declared in the declaration section.

• They should be explicitly raised in the Execution Section.

• They should be handled by referencing the user-defined exception name in the exception section.

For Example: Lets consider the product table and order_items table from sql joins to explain user-defined exception.

Lets create a business rule that if the total no of units of any particular product sold is more than 20, then it is a huge quantity and a special discount should be provided.

DECLARE 

  huge_quantity EXCEPTION; 

  CURSOR product_quantity is 

  SELECT p.product_name as name, sum(o.total_units) as units

  FROM order_tems o, product p

  WHERE o.product_id = p.product_id; 

  quantity order_tems.total_units%type; 

  up_limit CONSTANT order_tems.total_units%type := 20; 

  message VARCHAR2(50); 

BEGIN 

  FOR product_rec in product_quantity LOOP 

    quantity := product_rec.units;

     IF quantity > up_limit THEN 

      message := 'The number of units of product ' || product_rec.name ||  

                 ' is more than 20. Special discounts should be provided. 

		 Rest of the records are skipped. '

     RAISE huge_quantity; 

     ELSIF quantity < up_limit THEN 

      v_message:= 'The number of unit is below the discount limit.'; 

     END IF; 

     dbms_output.put_line (message); 

  END LOOP; 

 EXCEPTION 

   WHEN huge_quantity THEN 

     dbms_output.put_line (message); 

 END; 

/ 

RAISE_APPLICATION_ERROR ( )

RAISE_APPLICATION_ERROR is a built-in procedure in oracle which is used to display the user-defined error messages along with the error number whose range is in between -20000 and -20999.

Whenever a message is displayed using RAISE_APPLICATION_ERROR, all previous transactions which are not committed within the PL/SQL Block are rolled back automatically (i.e. change due to INSERT, UPDATE, or DELETE statements).

RAISE_APPLICATION_ERROR raises an exception but does not handle it.

RAISE_APPLICATION_ERROR is used for the following reasons,


a) to create a unique id for an user-defined exception.

b) to make the user-defined exception look like an Oracle error.

The General Syntax to use this procedure is:

RAISE_APPLICATION_ERROR (error_number, error_message); 

Steps to be folowed to use RAISE_APPLICATION_ERROR procedure:


1. Declare a user-defined exception in the declaration section.

2. Raise the user-defined exception based on a specific business rule in the execution section.

3. Finally, catch the exception and link the exception to a user-defined error number in RAISE_APPLICATION_ERROR.

Using the above example we can display a error message using RAISE_APPLICATION_ERROR.

DECLARE

  huge_quantity EXCEPTION; 

  CURSOR product_quantity is 

  SELECT p.product_name as name, sum(o.total_units) as units

  FROM order_tems o, product p

  WHERE o.product_id = p.product_id; 

  quantity order_tems.total_units%type; 

  up_limit CONSTANT order_tems.total_units%type := 20; 

  message VARCHAR2(50); 

BEGIN 

  FOR product_rec in product_quantity LOOP 

    quantity := product_rec.units;

     IF quantity > up_limit THEN 

        RAISE huge_quantity; 

     ELSIF quantity < up_limit THEN 

      v_message:= 'The number of unit is below the discount limit.'; 

     END IF; 

     Dbms_output.put_line (message); 

  END LOOP; 

 EXCEPTION 

   WHEN huge_quantity THEN 

	raise_application_error(-2100, 'The number of unit is above the discount limit.');

 END; 

/ 

How to use Explicit Cursor?

• DECLARE the cursor in the declaration section.
• OPEN the cursor in the Execution Section.
• FETCH the data from cursor into PL/SQL variables or records in the Execution Section.
• CLOSE the cursor in the Execution Section before you end the PL/SQL Block.

1) Declaring a Cursor in the Declaration Section:

   DECLARE

   CURSOR emp_cur IS 

   SELECT * 

   FROM emp_tbl

   WHERE salary > 5000; 

      In the above example we are creating a cursor ‘emp_cur’ on a query which returns the records of all the
      employees with salary greater than 5000. Here ‘emp_tbl’ in the table which contains records of all the
      employees.

2) Accessing the records in the cursor:
      Once the cursor is created in the declaration section we can access the cursor in the execution

      section of the PL/SQL program.

How to access an Explicit Cursor?

General Syntax to open a cursor is:

OPEN cursor_name;

General Syntax to fetch records from a cursor is:

FETCH cursor_name INTO record_name;

FETCH cursor_name INTO variable_list;

General Syntax to close a cursor is:

CLOSE cursor_name;

When a cursor is opened, the first row becomes the current row. When the data is fetched it is copied to the record or variables and the logical pointer moves to the next row and it becomes the current row. On every fetch statement, the pointer moves to the next row. If you want to fetch after the last row, the program will throw an error. When there is more than one row in a cursor we can use loops along with explicit cursor attributes to fetch all the records.

Points to remember while fetching a row:

· We can fetch the rows in a cursor to a PL/SQL Record or a list of variables created in the PL/SQL Block.
· If you are fetching a cursor to a PL/SQL Record, the record should have the same structure as the cursor.
· If you are fetching a cursor to a list of variables, the variables should be listed in the same order in the fetch statement as the columns are present in the cursor.

General Form of using an explicit cursor is:

 DECLARE

    variables;

    records;

    create a cursor;

 BEGIN 

   OPEN cursor;

   FETCH cursor;

     process the records;

   CLOSE cursor;

 END;

Lets Look at the example below


Example 1:

1> DECLARE 

2>    emp_rec emp_tbl%rowtype;

3>    CURSOR emp_cur IS 

4>    SELECT *

5>    FROM 

6>    WHERE salary > 10; 

7> BEGIN 

8>    OPEN emp_cur; 

9>    FETCH emp_cur INTO emp_rec; 

10>      dbms_output.put_line (emp_rec.first_name || '  ' || emp_rec.last_name); 

11>   CLOSE emp_cur; 

12> END; 

In the above example, first we are creating a record ‘emp_rec’ of the same structure as of table ‘emp_tbl’ in line no 2. We can also create a record with a cursor by replacing the table name with the cursor name. Second, we are declaring a cursor ‘emp_cur’ from a select query in line no 3 - 6. Third, we are opening the cursor in the execution section in line no 8. Fourth, we are fetching the cursor to the record in line no 9. Fifth, we are displaying the first_name and last_name of the employee in the record emp_rec in line no 10. Sixth, we are closing the cursor in line no 11.

What are Explicit Cursor Attributes?

Oracle provides some attributes known as Explicit Cursor Attributes to control the data processing while using cursors. We use these attributes to avoid errors while accessing cursors through OPEN, FETCH and CLOSE Statements.

When does an error occur while accessing an explicit cursor?

a) When we try to open a cursor which is not closed in the previous operation.
b) When we try to fetch a cursor after the last operation.

These are the attributes available to check the status of an explicit cursor.

Using Loops with Explicit Cursors:

Oracle provides three types of cursors namely SIMPLE LOOP, WHILE LOOP and FOR LOOP. These loops can be used to process multiple rows in the cursor. Here I will modify the same example for each loops to explain how to use loops with cursors.

Cursor with a Simple Loop:

1> DECLARE 

2>   CURSOR emp_cur IS 

3>   SELECT first_name, last_name, salary FROM emp_tbl; 

4>   emp_rec emp_cur%rowtype; 

5> BEGIN 

6>   IF NOT sales_cur%ISOPEN THEN 

7>      OPEN sales_cur; 

8>   END IF; 

9>   LOOP 

10>     FETCH emp_cur INTO emp_rec; 

11>     EXIT WHEN emp_cur%NOTFOUND; 

12>     dbms_output.put_line(emp_cur.first_name || ' ' ||emp_cur.last_name 

13>     || ' ' ||emp_cur.salary); 

14>  END LOOP; 

15>  END; 

16>  / 

In the above example we are using two cursor attributes %ISOPEN and %NOTFOUND.
In line no 6, we are using the cursor attribute %ISOPEN to check if the cursor is open, if the condition is true the program does not open the cursor again, it directly moves to line no 9.
In line no 11, we are using the cursor attribute %NOTFOUND to check whether the fetch returned any row. If there is no rows found the program would exit, a condition which exists when you fetch the cursor after the last row, if there is a row found the program continues.

We can use %FOUND in place of %NOTFOUND and vice versa. If we do so, we need to reverse the logic of the program. So use these attributes in appropriate instances.

Cursor with a While Loop:

Lets modify the above program to use while loop.

1> DECLARE 

2>  CURSOR emp_cur IS 

3>  SELECT first_name, last_name, salary FROM emp_tbl; 

4>  emp_rec emp_cur%rowtype; 

5> BEGIN 

6>   IF NOT sales_cur%ISOPEN THEN 

7>      OPEN sales_cur; 

8>   END IF; 

9>   FETCH sales_cur INTO sales_rec;  

10>  WHILE sales_cur%FOUND THEN  

11>  LOOP 

12>    dbms_output.put_line(emp_cur.first_name || ' ' ||emp_cur.last_name 

13>    || ' ' ||emp_cur.salary); 

15>    FETCH sales_cur INTO sales_rec; 

16>  END LOOP; 

17> END; 

18> / 

In the above example, in line no 10 we are using %FOUND to evaluate if the first fetch statement in line no 9 returned a row, if true the program moves into the while loop. In the loop we use fetch statement again (line no 15) to process the next row. If the fetch statement is not executed once before the while loop the while condition will return false in the first instance and the while loop is skipped. In the loop, before fetching the record again, always process the record retrieved by the first fetch statement, else you will skip the first row.

Cursor with a FOR Loop:

When using FOR LOOP you need not declare a record or variables to store the cursor values, need not open, fetch and close the cursor. These functions are accomplished by the FOR LOOP automatically.

General Syntax for using FOR LOOP:

FOR record_name IN cusror_name 

LOOP 

    process the row...

END LOOP; 

Let’s use the above example to learn how to use for loops in cursors.

1> DECLARE 

2>  CURSOR emp_cur IS 

3>  SELECT first_name, last_name, salary FROM emp_tbl; 

4>  emp_rec emp_cur%rowtype; 

5> BEGIN 

6>  FOR emp_rec in sales_cur 

7>  LOOP  

8>  dbms_output.put_line(emp_cur.first_name || ' ' ||emp_cur.last_name 

9>    || ' ' ||emp_cur.salary);  

10> END LOOP; 

11>END;

12> /

In the above example, when the FOR loop is processed a record ‘emp_rec’of structure ‘emp_cur’ gets created, the cursor is opened, the rows are fetched to the record ‘emp_rec’ and the cursor is closed after the last row is processed. By using FOR Loop in your program, you can reduce the number of lines in the program.

NOTE: In the examples given above, we are using backward slash ‘/’ at the end of the program. This indicates the oracle engine that the PL/SQL program has ended and it can begin processing the statements.

What are Cursors in PL/SQL

A cursor is a temporary work area created in the system memory when a SQL statement is executed. A cursor contains information on a select statement and the rows of data accessed by it. This temporary work area is used to store the data retrieved from the database, and manipulate this data. A cursor can hold more than one row, but can process only one row at a time. The set of rows the cursor holds is called the active set.

There are two types of cursors in PL/SQL:

Implicit cursors:

These are created by default when DML statements like, INSERT, UPDATE, and DELETE statements are executed. They are also created when a SELECT statement that returns just one row is executed.

Explicit cursors:

They must be created when you are executing a SELECT statement that returns more than one row. Even though the cursor stores multiple records, only one record can be processed at a time, which is called as current row. When you fetch a row the current row position moves to next row.

Implicit Cursors:

When you execute DML statements like DELETE, INSERT, UPDATE and SELECT statements, implicit statements are created to process these statements.

Oracle provides few attributes called as implicit cursor attributes to check the status of DML operations. The cursor attributes available are %FOUND, %NOTFOUND, %ROWCOUNT, and %ISOPEN.

For example, When you execute INSERT, UPDATE, or DELETE statements the cursor attributes tell us whether any rows are affected and how many have been affected.
When a SELECT... INTO statement is executed in a PL/SQL Block, implicit cursor attributes can be used to find out whether any row has been returned by the SELECT statement.
PL/SQL returns an error when no data is selected.

The status of the cursor for each of these attributes are defined in the below table. 

For Example: Consider the PL/SQL Block that uses implicit cursor attributes as shown below:

DECLARE  var_rows number(5);

BEGIN

  UPDATE employee 

  SET salary = salary + 1000;

  IF SQL%NOTFOUND THEN

    dbms_output.put_line('None of the salaries where updated');

  ELSIF SQL%FOUND THEN

    var_rows := SQL%ROWCOUNT;

    dbms_output.put_line('Salaries for ' || var_rows || 'employees are updated');

  END IF; 

END; 

In the above PL/SQL Block, the salaries of all the employees in the ‘employee’ table are updated. If none of the employee’s salary are updated we get a message 'None of the salaries where updated'. Else we get a message like for example, 'Salaries for 1000 employees are updated' if there are 1000 rows in ‘employee’ table.

]]>

What are Cursors in PL/SQL

A cursor is a temporary work area created in the system memory when a SQL statement is executed. A cursor contains information on a select statement and the rows of data accessed by it. This temporary work area is used to store the data retrieved from the database, and manipulate this data. A cursor can hold more than one row, but can process only one row at a time. The set of rows the cursor holds is called the active set.

There are two types of cursors in PL/SQL:

Implicit cursors:

These are created by default when DML statements like, INSERT, UPDATE, and DELETE statements are executed. They are also created when a SELECT statement that returns just one row is executed.

Explicit cursors:

They must be created when you are executing a SELECT statement that returns more than one row. Even though the cursor stores multiple records, only one record can be processed at a time, which is called as current row. When you fetch a row the current row position moves to next row.

Implicit Cursors:

When you execute DML statements like DELETE, INSERT, UPDATE and SELECT statements, implicit statements are created to process these statements.

Oracle provides few attributes called as implicit cursor attributes to check the status of DML operations. The cursor attributes available are %FOUND, %NOTFOUND, %ROWCOUNT, and %ISOPEN.

For example, When you execute INSERT, UPDATE, or DELETE statements the cursor attributes tell us whether any rows are affected and how many have been affected.
When a SELECT... INTO statement is executed in a PL/SQL Block, implicit cursor attributes can be used to find out whether any row has been returned by the SELECT statement.
PL/SQL returns an error when no data is selected.

The status of the cursor for each of these attributes are defined in the below table. 

For Example: Consider the PL/SQL Block that uses implicit cursor attributes as shown below:

DECLARE  var_rows number(5);

BEGIN

  UPDATE employee 

  SET salary = salary + 1000;

  IF SQL%NOTFOUND THEN

    dbms_output.put_line('None of the salaries where updated');

  ELSIF SQL%FOUND THEN

    var_rows := SQL%ROWCOUNT;

    dbms_output.put_line('Salaries for ' || var_rows || 'employees are updated');

  END IF; 

END; 

In the above PL/SQL Block, the salaries of all the employees in the ‘employee’ table are updated. If none of the employee’s salary are updated we get a message 'None of the salaries where updated'. Else we get a message like for example, 'Salaries for 1000 employees are updated' if there are 1000 rows in ‘employee’ table.

]]> http://naturebios.com/natureboard/showthread.php?tid=2718 Sat, 14 Nov 2009 15:16:20 -0800 http://naturebios.com/natureboard/showthread.php?tid=2718
There are three types of loops in PL/SQL:
• Simple Loop
• While Loop
• For Loop
1) Simple Loop

A Simple Loop is used when a set of statements is to be executed at least once before the loop terminates. An EXIT condition must be specified in the loop, otherwise the loop will get into an infinite number of iterations. When the EXIT condition is satisfied the process exits from the loop.
The General Syntax to write a Simple Loop is: