>>-PUT--+-cursor_id_var-+--------------------------------------->
        '-cursor_id-----'
 
>--+----------------------------------------------------+------><
   |       .-,----------------------------------------. |
   |       V                                          | |
   +-FROM----output_var--+--------------------------+-+-+
   |                     +-INDICATOR--indicator_var-+   |
   |                     +-$indicator_var-----------+   |
   |                     '-:indicator_var-----------'   |
   '-USING--+-SQL DESCRIPTOR--+-'descriptor '--+-+------'
            |                 '-descriptor_var-' |
            '-DESCRIPTOR--sqlda_pointer----------'
 

PUT stores a row in an insert buffer that is created when the cursor is opened.

If the buffer has no room for the new row when the statement executes, the buffered rows are written to the database in a block, and the buffer is emptied. As a result, some PUT statement executions cause rows to be written to the database, and some do not. You can use the FLUSH statement to write buffered rows to the database without adding a new row. The CLOSE statement writes any remaining rows before it closes an insert cursor.

If the current database uses explicit transactions, you must execute a PUT statement within a transaction.

The following example uses a PUT statement in ESQL/C:

EXEC SQL prepare ins_mcode from 
   'insert into manufact values(?,?)';
EXEC SQL declare mcode cursor for ins_mcode;
EXEC SQL open mcode;
EXEC SQL put mcode from :the_code, :the_name; 

The PUT statement is not an X/Open SQL statement. Therefore, you get a warning message if you compile a PUT statement in X/Open mode.

The values in the inserted row can come from one of the following sources:

• Constant values that are written into the INSERT statement
• Program variables that are named in the INSERT statement
• Program variables in the FROM clause of the PUT statement
• Values that are prepared in memory addressed by an sqlda structure or a system-descriptor area and then specified in the USING clause of the PUT statement

The system descriptor area or sqlda structure that descriptor or sqlda_pointer references must define a data type and memory location of each value that corresponds to a question-mark ( ? ) placeholder in a prepared INSERT statement.