Looping Statements

Loops allow you to execute a set of statements repeatedly until a certain condition is reached. The loop-terminating condition may be defined at the beginning, in the middle, or at the end of the loop using the WHILE LOOP, and REPEAT statements, respectively. Also, a FOR loop is available for iterating over a read-only result set and its terminating condition is when no more rows are left to read. Once the loop-terminating condition has been met, looping ceases and the flow of control continues on the line directly following the loop.


Variables cannot be declared within loops.


For readability, it is best to indent the loop body relative to the loop statement.

The WHILE and REPEAT loops are typically used when you do not know how many times to iterate through the loop prior to entering it. You should use the WHILE loop when you may not want to execute the loop even once, and the REPEAT loop when you want to ensure that the statements within the loop are executed at least once. The FOR loop is used for situations where you need to iterate over a read-only result set, using result set values for some purpose such as defining the value of a variable. LOOP is generally used if you have multiple exit conditions for the loop, perhaps at various possible locations.

FOR Loop

The FOR loop statement is supported on DB2 for LUW and iSeries only. It is used to iterate over a read-only result set that is defined by its select-statement. Looping will cease when there are no rows left in the result set. You can easily use a WHILE loop to achieve the same result in DB2 for zSeries.


Positioned updates and deletes are not supported in the FOR loop. However, searched updates and deletes are allowed.

The syntax is depicted in Figure 4.12.

Figure 4.12. A FOR statement syntax diagram.
 >>-+--------+--FOR--for-loop-name--AS--------------------------->    '-label:-' >--+----------------------------------------------+------------->    |                                              |    '-cursor-name--CURSOR--+-----------+--FOR------'                           '-WITH HOLD-'                          .----------------------------.                          V                            | >--select-statement--DO----SQL-procedure-statement--;-+---------> >--END FOR--+-------+------------------------------------------><             '-label-' 

The for-loop-name specifies a label for the implicit compound statement generated to implement the FOR statement. It follows the rules for the label of a compound statement. The for-loop-name can used to qualify the column names in the result set as returned by the select-statement.

The cursor-name simply names the cursor that is used to select rows from the result set. If not specified, DB2 will automatically generate a unique cursor name internally.

The WITH HOLD option dictates the behavior of cursors across multiple units of work. It is described in detail in Chapter 5, "Understanding and Using Cursors and Result Sets."

The column names of the select-statement must be unique, and a FROM clause specifying a table (or multiple tables if doing some kind of JOIN or UNION) is required. The table(s) and column(s) referenced must exist prior to the loop being executed. This allows you to iterate over result sets that are formed from tables which exist prior to invoking the stored procedure, or tables that have been created by a previous SQL PL statement (such as declared user-temporary tables, which are discussed in Chapter 10, "Leveraging DB2 Application Development Features").

The FOR loop is essentially a CURSOR defined by the select-statement. This CURSOR cannot be referenced outside of the FOR loop, however, so OPEN, FETCH, and CLOSE statements will result in error.

In Figure 4.13, the FOR loop is used to iterate over all rows of the employee table (because no WHERE clause is being used). For each row, it generates a full name from the values out of three columnslast name, first name, and middle name initial.

Figure 4.13. A FOR loop example.
 -- Example applies to LUW and iSeries only --   because of the usage of FOR loops. CREATE PROCEDURE demo_for_loop ( )     LANGUAGE SQL     SPECIFIC demo_for_loop                          -- applies to LUW and iSeries dfl: BEGIN     -- Declare variables     DECLARE v_fullname VARCHAR(50);     -- Procedure logic     FOR v_row AS SELECT firstnme, midinit, lastname                -- (1)                    FROM employee       DO       SET v_fullname = v_row.lastname || ', ' ||                        v_row.firstnme || ' ' || v_row.midinit;     -- (2)       INSERT INTO tname VALUES (v_fullname);                       -- (3)    END FOR; END dfl 

You can see the defining select-statement on Line (1) and where the columns of the result set are being concatenated together to form the v_fullname on Line (2). Finally, this newly formed v_fullname is inserted into a table called tname on Line (3).

To test the procedure in Figure 4.13, you need to have a table created as

 CREATE TABLE tname ( fullname VARCHAR(50) ) 


Please note that the simple example in Figure 4.13 is only used to show an example on how to use the FOR loop to go through a result set from a SELECT statement. The logic of the example is so simple that you can easily accomplish the same with one SQL statement instead. However, you should realize that the FOR loop can accommodate more complex programming logic than a simple SQL statement can, and that is when you should consider using the FOR loop in your own applications.


The defining feature of a WHILE loop is that its looping condition is evaluated prior to initial loop execution and all following loop iterations. The WHILE loop will continue to execute until the looping condition evaluates to false.


Be sure not to define a condition that always evaluates to true, or you will become caught in an infinite loop.

When defining the looping condition, be sure to specify a full conditional statement (that includes operators). Otherwise, your SQL procedure will not build. For example,

 WHILE (variable) DO         statement1;         statement2; END WHILE; 

is not enough. You need to use an operator, as in

 WHILE (variable = 1) DO         statement1;         statement2; END WHILE; 

The syntax for the WHILE loop is illustrated in Figure 4.14.

Figure 4.14. A WHILE loop syntax diagram.
 >>-+---------+--WHILE--search-condition--DO--------------------->    '-:label--'      .-------------------------------.      V                               | >--------SQL-procedure-statement--;---+--END WHILE--------------> >-----+--------+-----------------------------------------------><       '-label--' 

The search-condition specifies a condition that is evaluated before each execution of the loop. If the condition is true, the SQL-procedure-statements in the loop are processed.

Figure 4.15 illustrates how to use a WHILE loop to sum all integer values between n and m (which are assumed to be positive and provided by input parameters to the procedure).

Figure 4.15. A simple WHILE loop example.
 CREATE PROCEDURE sum_mn ( IN p_start INT                         , IN p_end INT                         , OUT p_sum INT )     LANGUAGE SQL     SPECIFIC sum_mn                                -- applies to LUW and iSeries  -- WLM ENVIRONMENT <env>                           -- applies to zSeries smn: BEGIN    DECLARE v_temp INTEGER DEFAULT 0;    DECLARE v_current INTEGER;    SET v_current = p_start;    WHILE (v_current <= p_end) DO        SET v_temp = v_temp + v_current;            -- (1)        SET v_current = v_current + 1;    END WHILE;    SET p_sum = v_temp; END smn 

This example is fairly simple and is intended to show you how the WHILE loop works using as little code as possible. More commonly, however, a WHILE loop is used to repeatedly perform SQL procedure statements, such as FETCH (for retrieving row values from a cursor). For examples of using WHILE loops with cursor operations such as OPEN, FETCH and CLOSE, see Chapter 5, "Understanding and Using Cursors and Result Sets."


In the WHILE loop, you saw that the looping condition is evaluated at the very beginning of the loop. If the looping condition evaluates to false at this first examination, then the loop body will not execute at all.

In some cases, however, it may be necessary that the loop be executed at least once. This is where the REPEAT loop is useful. A REPEAT loop ensures that at least one iteration of the loop is completed. This is the case because the looping condition is not evaluated until the final line of code in the loop.

The syntax for the REPEAT loop is shown in Figure 4.16.

Figure 4.16. A REPEAT loop syntax diagram.
                           .-------------------------------.                           V                               | >>-+---------+--REPEAT-------SQL-procedure-statement--;---+----->    '-label:--' >----UNTIL--search-condition---END REPEAT----+--------+--------><                                              '-label--' 

In Figure 4.17, the procedure in Figure 4.15 is re-implemented using REPEAT.

Figure 4.17. A REPEAT loop example.
 CREATE PROCEDURE sum_mn2 ( IN p_start INT                          , IN p_end INT                          , OUT p_sum INT )     LANGUAGE SQL     SPECIFIC sum_mn2                             -- applies to LUW and iSeries  -- WLM ENVIRONMENT <env>                        -- applies to zSeries smn2: BEGIN    DECLARE v_temp INTEGER DEFAULT 0;    DECLARE v_current INTEGER;    SET v_current = p_start;    REPEAT       SET v_temp = v_temp + v_current;                  -- (1)       SET v_current = v_current + 1;    UNTIL (v_current > p_end)    END REPEAT;    SET p_sum = v_temp; END smn2 

If you execute both the WHILE loop example in Figure 4.15 and the REPEAT loop example in Figure 4.17 with p_start smaller than p_end, you will see the difference between the two. The WHILE loop example in Figure 4.15 generates p_sum with 0. The statement on Line (1) in Figure 4.15 is never executed because the condition in the WHILE clause is not met. The REPEAT loop example generates p_sum with the value of p_start instead because the statement on Line (1) in Figure 4.17 is always executed once.


The LOOP statement is somewhat different from the other types of loops that you have seen thus far. The LOOP does not have a terminating condition clause that is part of its declaration statement. It will continue to loop until some other piece of code inside it explicitly forces the flow of control to jump to some point outside of the loop.

LOOP will commonly have some logic that eventually branches to a LEAVE statement. You can also use a GOTO statement instead of a LEAVE, but the use of GOTO is discouraged.


Ensure that some action within the loop eventually invokes a LEAVE or GOTO statement. Otherwise, your code can become caught in an infinite loop.

The LOOP syntax is illustrated in Figure 4.18.

Figure 4.18. A LOOP syntax diagram.
                         .-------------------------------.                         V                               | >>-+---------+--LOOP-------SQL-procedure-statement--;---+------->    '-label:--' >----END LOOP----+--------+------------------------------------><                  '-label--' 


There is no terminating condition defined within the LOOP syntax itself.

An example of using LOOP is deferred until the discussion on LEAVE in the next section.

    DB2 SQL PL. Deployment and Advanced Configuration Essential Guide for DB2 UDB on Linux., UNIX, Windows, i5. OS, z. OS
    Intelligent Enterprises of the 21st Century
    ISBN: N/A
    EAN: 2147483647
    Year: 2003
    Pages: 205

    flylib.com © 2008-2017.
    If you may any questions please contact us: flylib@qtcs.net