Performance Tuning in SAP ABAP Performance tuning techniques in SAP
Run-time analysis Check efficiency of SAP ABAP programs and Function modules using SAP run-time analysis
Run-time analysis is used to check the efficiency of a program or function module or t-
code in terms of what is the load on the database server, application server, presentation server etc. The run-time analysis will display the load in a graph with %'s and the time in micro
seconds. The graph will be displayed with either red color or green color. If the graph contains green color, then the program execution time is very good or very less. If the graph contains red color, the program execution time is very bad or very long . Always make sure the the load on database server should be less than 40% and should be green.
T-code for run-time analysis is SE30 or SAT(Latest Versions).
Example program on run time analysys in SAP ABAP Program1: Create a program to get MARA details and display as below. REPORT ZSAN_SE30. **Runtime analysis using SE30 DATA : IT_MARA TYPE TABLE OF MARA. DATA : WA_MARA TYPE MARA. SELECT * FROM MARA INTO TABLE IT_MARA UP TO 500 ROWS.
LOOP AT IT_MARA INTO WA_MARA. WRITE:/ WA_MARA-MATNR, WA_MARA-MTART, WA_MARA-MBRSH, WA_MARA-MATKL, WA_MARAMEINS. ENDLOOP.
Go to SE30, provide program name, click on execute button, the out put will bi displayed, click on back and click on evaluate to see run-time analysis.
The result will be like below(may vary based on server capacity).
In the above image you can see ABAP and database layers are in red color, means the time taken to process these layers is more. Program:2 Now change the above program like below and test again. REPORT ZSAN_SE30. TYPES: BEGIN OF TY_MARA, MATNR TYPE MARA-MATNR,
MTART TYPE MARA-MTART, MBRSH TYPE MARA-MBRSH, MATKL TYPE MARA-MATKL, MEINS TYPE MARA-MEINS, END OF TY_MARA. **Runtime analysis using SE30 DATA : IT_MARA TYPE TABLE OF TY_MARA. DATA : WA_MARA TYPE ty_MARA. SELECT MATNR MTART MBRSH MATKL MEINS FROM MARA INTO TABLE IT_MARA UP TO 500 ROWS. LOOP AT IT_MARA INTO WA_MARA. WRITE:/ WA_MARA-MATNR, WA_MARA-MTART, WA_MARA-MBRSH, WA_MARA-MATKL, WA_MARA-MEINS. ENDLOOP.
Now go to SE30 and follow the above process.
Now the evaluation will be like below.
You can see the database layer is in green color means the standards are good. A good program with coding standards always have database layers is in green color.
Performance tuning ST05 SAP performance tuning to check the performance of SAP ABAP programs ST05 is a T-code which is used to trace performance of SAP Objects, things we can done using ST05 are below.
SQL Trace which is used to trace time execution for open SQL statements like select, insert, update, delete etc. Enqueue Trace which is used to trace locking and unlocking of lock objects. Buffer Trace, this is used to monitor table buffer accesses. RFC Trace, this is used to trace remote calls made of a application. HTTP Trace which is used to trace HTTP requests(ex: browser requests) of SAP applications.
To perform trace operations go to ST05, select trace, activate trace, open another session and execute the application which we wants to trace, comeback to ST05, deactivate the trace and display the trace.
ABAP programing standards for performance tuning Programing standards for SAP ABAP programers, best practices SAP ABAP programing Programming Standard
Never use select *, always use select with list of fields because if you use select * it will get all columns from database, if you use select list of fields it will get selected columns only.
Always specify key fields in where condition, because it will be having primary index.
Bad example
SELECT * FROM MARA INTO TABLE ...
SELECT * FROM MARA INTO TABLE IT_MARA
Good example
SELECT MATNR MTART MEINS MBRSH FROM MARA ..
SELECT MATNR MTART MEINS FROM MARA INTO TABLE IT_MARA
WHERE MTART = 'FERT'.
WHERE MATNR = '0001'.
Some times we may need to use non-key fields in where condition, in such cases create secondary index to improve performance. Never use select with corresponding .
SELECT * FROM MARA INTO CORRESPONDING FIELDS OF TABLE IT_MARA.
Always use read table key with READ TABLE IT_MARA INTO binary search, binary search WA_MARA WITH KEY MATNR = will improve performance. '001'
SELECT MATNR MTART MEINS FROM MARA INTO TABLE IT_MARA.
READ TABLE IT_MARA INTO WA_MARA WITH KEY MATNR = '001'.
BINARY SEARCH.
Avoid using select joins for more than 3 database tables.
Select Joins in SAP ABAP
Using select for all entries
Always use select for all entries for more than three tables.
Select Joins in SAP ABAP
Using select for all entries
Programming Standard
Bad example
When using select for all SELECT * FROM MAKT entries, check whether the parent internal table is initial or INTO IT_MAKT not .
FOR ALL ENTRIES IN IT_MARA
Never use nested loops.
Good example
IF IT_MARA IS NOT INITIAL. SELECT * FROM MAKT INTO IT_MAKT
FOR ALL ENTRIES IN IT_MARA
WHERE MATNR = IT_MARA-MATNR.
WHERE MATNR = IT_MARA-MATNR.
ENDIF.
LOOP AT IT_MARA INTO WA_MARA.
LOOP AT IT_MARA INTO WA_MARA.
READ TABLE IT_MAKT INTO WA_MAKT WHERE MATNR = WA_MARAMATNR.
ENDLOOP.
LOOP AT IT_MAKT INTO WA_MAKT.
ENDLOOP.
ENDLOOP.
Never use select statements inside loops .
LOOP AT IT_MARA INTO INSTEAD USE FOR ALL WA_MARA.
SELECT * FROM ENTRIES MAKT ...
ENDLOOP.
Increase performance using Parallel Cursor in SAP ABAP The first thing that comes into the mind of a good ABAPer is performance of his object, the main problem for performance issues are bad Select statements ex: select joins for more than 3 tables, nested Loops, how ever there is a concept of parallel cursor in SAP ABAP to avoid performance issues with nested loops with where condition. To understand parallel cursor, follow and understand the below programs.
Program without parallel cursor REPORT ZSAN_PARALLEL_CURSER. TYPES: BEGIN OF TY_MARA, MATNR TYPE MARA-MATNR, MTART TYPE MARA-MTART, MBRSH TYPE MARA-MBRSH, MATKL TYPE MARA-MATKL, MEINS TYPE MARA-MEINS, END OF TY_MARA. DATA : IT_MARA TYPE TABLE OF TY_MARA, WA_MARA TYPE TY_MARA. TYPES: BEGIN OF TY_MVKE, MATNR TYPE MVKE-MATNR, VKORG TYPE MVKE-VKORG, VTWEG TYPE MVKE-VTWEG, END OF TY_MVKE. DATA : IT_MVKE TYPE TABLE OF TY_MVKE, WA_MVKE TYPE TY_MVKE. SELECT MATNR MTART MBRSH MATKL MEINS FROM MARA INTO TABLE IT_MARA UP TO 500 ROWS. SORT IT_MARA. DELETE ADJACENT DUPLICATES FROM IT_MARA COMPARING ALL FIELDS. SELECT MATNR
VKORG VTWEG FROM MVKE INTO TABLE IT_MVKE FOR ALL ENTRIES IN IT_MARA WHERE MATNR = IT_MARA-MATNR. LOOP AT IT_MARA INTO WA_MARA. LOOP AT IT_MVKE INTO WA_MVKE WHERE MATNR = WA_MARA-MATNR. WRITE:/ WA_MARA-MATNR, WA_MARA-MTART, WA_MARA-MATKL, WA_MVKE-VKORG, WA_MVKEVTWEG. ENDLOOP. ENDLOOP.
Program with parallel cursor. REPORT ZSAN_PARALLEL_CURSER. TYPES: BEGIN OF TY_MARA, MATNR TYPE MARA-MATNR, MTART TYPE MARA-MTART, MBRSH TYPE MARA-MBRSH, MATKL TYPE MARA-MATKL, MEINS TYPE MARA-MEINS, END OF TY_MARA. DATA : IT_MARA TYPE TABLE OF TY_MARA, WA_MARA TYPE TY_MARA. TYPES: BEGIN OF TY_MVKE, MATNR TYPE MVKE-MATNR, VKORG TYPE MVKE-VKORG, VTWEG TYPE MVKE-VTWEG, END OF TY_MVKE. DATA : IT_MVKE TYPE TABLE OF TY_MVKE, WA_MVKE TYPE TY_MVKE. SELECT MATNR MTART MBRSH MATKL MEINS FROM MARA INTO TABLE IT_MARA UP TO 500 ROWS. SORT IT_MARA. DELETE ADJACENT DUPLICATES FROM IT_MARA COMPARING ALL FIELDS.
SELECT MATNR VKORG VTWEG FROM MVKE INTO TABLE IT_MVKE FOR ALL ENTRIES IN IT_MARA WHERE MATNR = IT_MARA-MATNR. SORT: IT_MARA, IT_MVKE. DATA V_INDEX TYPE SY-TABIX. LOOP AT IT_MARA INTO WA_MARA. READ TABLE IT_MVKE INTO WA_MVKE WITH KEY MATNR = WA_MARA-MATNR BINARY SEARCH. IF SY-SUBRC = 0. V_INDEX = SY-TABIX. LOOP AT IT_MVKE INTO WA_MVKE FROM V_INDEX. "no where condition for loop IF WA_MVKE-MATNR <> WA_MARA-MATNR. EXIT. ENDIF. WRITE:/ WA_MARA-MATNR, WA_MARA-MTART, WA_MARA-MATKL, WA_MVKE-VKORG, WA_MVKEVTWEG. ENDLOOP. ENDIF. CLEAR: WA_MARA, WA_MVKE. ENDLOOP.
10 best ABAP programming standards for performance tuning Below are the most common and best programming standards for developing ABAP applications.
1. When reading data from database table, never use SELECT *, always use select with list of fields. 2. Always specify key fields in where conditions of SELECT statements, because these will have primary index. 3. Sometimes we may need to use non-key fields in where conditions of SELECT statements, in that case create secondary indexes and use. 4. Never use select with corresponding in SELECT statements. 5. Always use read table with binary search and make you have sorted the internal table in ascending order before using read table binary search. 6. Never use select joins for more than 3 tables. 7. Always use select for all entries for more than 3 tables. 8. Always check whether the parent internal table is initial or not before using for all entries. 9. Never use nested loops, instead use parallel cursor . 10. Never use select statements inside loops, instead use for all entries.
