SQL Cursor Commands

The following ALLBASE/SQL commands are used in cursor processing:

DECLARE CURSOR defines a cursor and associates it with a query.
OPEN defines the active set.
FETCH retrieves one row of the active set into host variables; when a row resides in host variables it is known as the current row. When a row is current and the active set is a query result derived from a single table, you can use one of the following two commands to change the row.
UPDATE WHERE CURRENT updates the current row.
DELETE WHERE CURRENT deletes the current row.
CLOSE terminates access to the active set and frees up ALLBASE/SQL buffer space used to handle the cursor.

For a given cursor, the commands listed above (with the exception of DECLARE CURSOR) should be contained within the same transaction. Refer to the ALLBASE/SQL Reference Manual for the complete syntax and semantics of these commands.

DECLARE CURSOR

The DECLARE CURSOR command names a cursor and associates it with a particular SELECT command:

   DECLARE CursorName

           [IN DBEFileSetName]

           CURSOR FOR

           SelectCommand

           [FOR UPDATE OF ColumnName [,ColumnName...]

Note that the DECLARE CURSOR command has two optional clauses:

The IN clause defines the DBEFileSet in which the section generated by the preprocessor for this command is stored. If no IN clause is specified, file space in the SYSTEM DBEFileSet is used.
The FOR UPDATE OF clause is used when you intend to use the UPDATE WHERE CURRENT command to update a current row. This command may offer the simplest way to update a current row, but it imposes certain restrictions on the SELECT command. Updating a current row is discussed fully later in this chapter under "Update Where Current."

The SELECT command for cursor declarations that do not include the FOR UPDATE clause can consist of any of the SELECT command clause except the INTO clause:

     SELECT SelectList

       FROM TableNames

      WHERE SearchCondition1

   GROUP BY ColumnNames

     HAVING SearchCondition2

   ORDER BY ColumnIdentifiers

A SELECT command associated with a cursor does not name output host variables, but may name input host variables in the select list, the WHERE clause, and the HAVING clause. In the following example, the rows qualifying for the query result will be those with a CountCycle matching that specified by the user in input host variable CountCycle:

   EXEC SQL DECLARE Inventory

             CURSOR FOR

             SELECT PartNumber,

                    BinNumber,

                    QtyOnHand,

                    AdjustmentQty

               FROM PurchDB.Inventory

              WHERE CountCycle = :CountCycle

           ORDER BY BinNumber;

When performing cursor processing, the ORDER BY clause may be useful. In the previous example, the rows in the query result will be in order by ascending bin number to help the program user, who will be moving from bin to bin, taking a physical inventory.

The DECLARE CURSOR command is actually a preprocessor directive. When the preprocessor parses this command, it stores a section in the target DBEnvironment. At run time, the section is not executed when the DECLARE CURSOR command is encountered. The section is executed when the OPEN command is encountered. Because the DECLARE CURSOR command is not executed at run time, you do not need to perform error status checking in your program following this command.

OPEN

The OPEN command allocates internal buffer space and defines the active set:

OPEN CursorName [KEEP CURSOR [[WITH LOCKS] [NOLOCKS]] ]

The following command opens the cursor defined earlier:

    EXEC SQL OPEN Inventory;

Once the active set is defined, the FETCH command will retrieve data from it, one row at a time.

You can use the KEEP CURSOR WITH NOLOCKS option for a cursor that involves sorting, whether through the use of a DISTINCT, GROUP BY, or ORDER BY clause, or as the result of a union or a join operation. However, for kept cursors involving sorting, ALLBASE/SQL does not ensure data integrity.

It is your responsibility to ensure data integrity by verifying the continued existence of a row before updating it or using it as the basis for updating some other table. For an updatable cursor, you can use either the REFETCH or SELECT command to verify the continued existence of a row. For a cursor that is non-updatable, you must use the SELECT command.

A warning (DBWARN 2056) regarding the kept cursor on a sort with no locks is generated. You must check for this warning if you want to detect the execution of this type of cursor operation.

FETCH

The FETCH command defines a current row and delivers the row into output host variables:

   FETCH CursorName INTO OutputHostVariables

Remember to include indicator variables when one or more columns in the query result may contain a null value, for example:

   EXEC SQL FETCH  Inventory

             INTO :PartNumber,

                  :BinNumber,

                  :QtyOnHand     :QtyOnHandInd,

                  :AdjustmentQty :AdjustmentQtyInd;

The first time you execute the FETCH command, the first row in the query result becomes the current row. With each subsequent execution of the FETCH command, each succeeding row in the query result becomes the current row. After the last row in the query result has been fetched, ALLBASE/SQL sets sqlca.sqlcode to 100. ALLBASE/SQL also sets sqlca.sqlcode to 100 if no rows qualify for the active set. You should test for an sqlca.sqlcode value of 100 after each execution of the FETCH command to determine whether to re-execute the command.

   #define   TRUE    0

   #define   FALSE   1

     .

     .

     .

   int GetARow()

   {

   int  DoFetch;

   .

   .

   .

   DoFetch = TRUE;

   do {

   .

   .     The FETCH command appears here.

   .

   switch (sqlca.sqlcode) {

     case 0:           DisplayRow();

                       break;

     case 100:         DoFetch = FALSE;

                       CloseCursor();

                       CommitWork();

                       break;

     default:          DoFetch = FALSE;

                       SQLStatusCheck();

                       CloseCursor();

                       RollBack();

                       break;

                       }

     } while (DoFetch == TRUE);

     }

When a row is current, you can update it by using the UPDATE WHERE CURRENT command or delete it by using the DELETE WHERE CURRENT command.

UPDATE WHERE CURRENT

This command can be used to update the current row when the SELECT command associated with the cursor does not contain one of the following:

DISTINCT clause in the select list.
Aggregate function in the select list.
FROM clause with more than one table.
ORDER BY clause.
GROUP BY clause.

The UPDATE WHERE CURRENT command identifies the active set to be updated by naming the cursor and the column(s) to be updated:

   UPDATE TableName

      SET ColumnName = ColumnValue [,...]

    WHERE CURRENT OF CursorName

Any columns you name in this command must also have been named in a FOR UPDATE clause in the related DECLARE CURSOR command, for example:

   EXEC SQL DECLARE AdjustQtyOnHand

             CURSOR FOR

             SELECT PartNumber,

                    BinNumber,

                    QtyOnHand,

                    AdjustmentQty

               FROM PurchDB.Inventory

              WHERE QtyOnHand IS NOT NULL

                AND AdjustmentQty IS NOT NULL

      FOR UPDATE OF QtyOnHand,

                    AdjustmentQty;



   EXEC SQL OPEN AdjustQtyOnHand;

   .

   In this case, the output host variables do not need to include

   indicator variables, because the SELECT command associated with the

   cursor eliminates any rows having null values from the active set.

   .

   EXEC SQL FETCH  AdjustQtyOnHand

             INTO :PartNumber,

                  :BinNumber,

                  :QtyOnHand,

                  :AdjustmentQty;

   .

   .

   EXEC SQL UPDATE PurchDB.Inventory

               SET QtyOnHand     = :QtyOnHand + :AdjustmentQty,

                   AdjustmentQty = 0

             WHERE CURRENT OF AdjustQtyOnHand;

In the previous example, the order of the rows in the query result is not important. Therefore the SELECT command associated with the cursor AdjustQtyOnHand does not need to contain an ORDER BY clause and the UPDATE WHERE CURRENT command can be used.

In cases where order is important and the ORDER BY clause must be used, you can use the UPDATE command with the WHERE clause to update values in the current row, as well as other rows that qualify for the search condition:

   EXEC SQL DECLARE Inventory

             CURSOR FOR

             SELECT PartNumber,

                    BinNumber,

                    QtyOnHand,

                    AdjustmentQty

               FROM PurchDB.Inventory

              WHERE CountCycle = :CountCycle

           ORDER BY BinNumber;

   .

   .

   .

   EXEC SQL FETCH  Inventory

             INTO :PartNumber,

                  :BinNumber,

                  :QtyOnHand     :QtyOnHandInd

                  :AdjustmentQty :AdjustmentQtyInd;

   .

   The program displays the current row.  If the QtyOnHand value is not null,

   the program prompts the user for an adjustment quantity.  Adjustment

   quantity is the difference between the quantity actually in the bin and the

   QtyOnHand in the row displayed.  If the QtyOnHand value is null, the program

   prompts the user for both QtyOnHand and AdjustmentQty.  Any value entered is

   used in the following UPDATE command.

   .

   .

   EXEC SQL UPDATE PurchDB.Inventory

               SET QtyOnHand =     :QtyOnHand :QtyOnHandInd,

                   AdjustmentQty = :AdjustmentQty :AdjustmentQtyInd

             WHERE PartNumber = :PartNumber

               AND BinNumber  = :BinNumber;

After either the UPDATE WHERE CURRENT or the UPDATE command is executed, the current row remains the same until the FETCH command is re-executed.

If you want to execute UPDATE commands inside the FETCH loop, remember that more than one row in the active set may qualify for the UPDATE operation, as when the WHERE clause in the the UPDATE command does not specify a unique key. When more than one row qualifies for the UPDATE, you may not see a changed row unless you CLOSE and re-OPEN the cursor. To avoid this problem, either ensure that your UPDATE commands change only one row (the current row) or perform the UPDATE operations outside the FETCH loop.

DELETE WHERE CURRENT

This command can be used to delete the current row when the SELECT command associated with the cursor does not contain one of the following:

DISTINCT clause in the select list.
Aggregate function in the select list.
FROM clause with more than one table.
ORDER BY clause.
GROUP BY clause.

The DELETE WHERE CURRENT command has a very simple structure:

   DELETE FROM TableName

         WHERE CURRENT OF CursorName

The DELETE WHERE CURRENT command can be used in conjunction with a cursor declared with or without the FOR UPDATE clause:

   The program displays the current row and asks the user whether to update or

   delete it.  If the user wants to delete the row, the following command

   is executed.



   EXEC SQL DELETE FROM PurchDB.Inventory

             WHERE CURRENT OF AdjustQtyOnHand;

Even though the SELECT command associated with cursor Inventory names only some of the columns in table PurchDB.Inventory, the entire current row is deleted.

After the DELETE WHERE CURRENT command is executed, there is no current row. You must re-execute the FETCH command to obtain another current row.

As with the UPDATE WHERE CURRENT command, if the SELECT command associated with the cursor contains an ORDER BY clause or other components listed earlier, you can use the DELETE command with the WHERE clause to delete a row:

   EXEC SQL DELETE FROM PurchDB.Inventory

                  WHERE PartNumber = :PartNumber

                    AND BinNumber  = :BinNumber;

If you use the DELETE command to delete a row while using a cursor to examine an active set, remember that more than one row will be deleted if multiple rows satisfy the conditions specified in the WHERE clause of the DELETE command. In addition, the row that is current when the DELETE command is executed remains the current row until the FETCH command is re-executed.

CLOSE

When you no longer want to operate on the active set, you use the CLOSE command:

   CLOSE CursorName

The CLOSE command frees up ALLBASE/SQL internal buffers used to handle cursor operations. This command does not release any locks obtained since the cursor was opened; to release locks, you must terminate the transaction with a COMMIT WORK or a ROLLBACK WORK:

   The program opens a cursor and operates on the active set.  After the last

   row has been operated on, the cursor is closed.



   EXEC SQL CLOSE Inventory;



   Additional SQL commands are executed, then the transaction is terminated.



   EXEC SQL COMMIT WORK;

You also use the CLOSE command when you want to re-access the active set. In this case, simply re-open the cursor after executing the CLOSE command. Because locks have not been released, any changes to the rows in the active set will be those made by your program since the cursor was first opened:

   Cursor Inventory is used to update information

   in table PurchDB.Inventory.  After the last row

   in the active set has been fetched and its information

   changed, the cursor is closed.



   EXEC SQL CLOSE Inventory;



   The cursor is then re-opened to allow the program

   user to review the information and optionally make

   last-minute adjustments.



   EXEC SQL OPEN Inventory;



   After the user has reviewed all rows in the active

   set, any changes made to the active set are

   made permanent as follows.



   EXEC SQL COMMIT WORK;

When a transaction terminates, any cursors opened during that transaction are automatically closed unless you are using the KEEP CURSOR option of the OPEN command. To avoid possible confusion, it is good programming practice to always use the CLOSE command followed by COMMIT WORK to explicitly close any open cursors before ending a transaction.