Performance Tuning Example 2(using a 'For All Entries' statement )

Source code below is an example of using the 'For All Entries' statement

REPORT ZHIS_TUN_002_BEFORE.

DATA: IT_COEP TYPE TABLE OF COEP,

           IT_COBK TYPE TABLE OF COBK.

SELECT * INTO CORRESPONDING FIELDS OF TABLE IT_COBK

FROM COBK.

SELECT *

FROM  COEP INTO CORRESPONDING FIELDS OF TABLE IT_COEP

             FOR ALL ENTRIES IN IT_COBK

WHERE KOKRS = IT_COBK-KOKRS

AND BELNR = IT_COBK-BELNR.

WRITE:/ LINES( IT_COEP ).

The results of analysis of the source code of the above in RUNTIME ANALYSIS are as follows:

It took most of the work time in the ‘DB: OPEN’ and ‘DB: Fetch’

but 'DB: Fetch' time can not improve because the amount of data transferred is large.

Double-click the ‘DB: OPEN’, check the HIT LIST

If the data is large, 'For All Entries' may be generated many DB access.

As a result, It can be confirmed that the execution time is slow

The following is the source code that uses the JOIN in order to improve this.

REPORT ZHIS_TUN_002_AFTER.

DATA: IT_COEP TYPE TABLE OF COEP ,

           IT_COBK TYPE TABLE OF COBK .

SELECT * INTO CORRESPONDING FIELDS OF TABLE IT_COEP

FROM  COEP AS A INNER JOIN COBK AS B

                        ON B~KOKRS = A~KOKRS AND B~BELNR = A~BELNR.

WRITE:/ LINES( IT_COEP ) .

Because there was a 'COEP' table only once access, you can see that the execution time is greatly reduced.

(DB:OPEN Runtime : 14,103,513 -> 752)

If the data is large , do not the 'For All Entries' and Use 'JOIN' or 'SUBQUERY' can improve performance.