topic Re: Issue in Application Development and Automation Discussions

Issue

Former Member — Fri, 10 Nov 2006 22:00:36 GMT

I have issue and need to solve urgenlty.

The program is dumping here. How can I solve this issue.

Could you please help me.

SELECT BUKRS BELNR GJAHR BUZEI LIFNR SGTXT MWSKZ HKONT MENGE WRBTR

EBELN EBELP PRCTR WERKS SHKZG ZEKKN NPLNR KOSTL PROJK MATNR

AUFPL APLZL MEINS

FROM BSEG

INTO TABLE INTBSEG

WHERE BUKRS = TBKPF-BUKRS

AND BELNR = TBKPF-BELNR

AND GJAHR = TBKPF-GJAHR

AND MWART NE 'V'

AND TXJCD IN TXJCODE.

Re: Issue

Former Member — Fri, 10 Nov 2006 22:02:20 GMT

Hi,

Check the sequence of the fields declared in the internal table and given in the select statement..

If you were not able to find out..Use INTO CORRESPONDING FIELDS OF TABLE instead of INTO TABLE..If the field names are same..

ALso Give us the description of the SHORT DUMP..

Thanks,

Naren

Re: Issue

Former Member — Fri, 10 Nov 2006 22:05:36 GMT

The SELECT looks OK. Are you sure that all of the fields you are selecting are in your internal table? In the same order?

Rob

Re: Issue

Former Member — Fri, 10 Nov 2006 22:08:01 GMT

What happened?

The program "ZKRAR232" has exceeded the maximum permitted runtime without

interruption, and has therefore been terminated.

Error analysis

After a certain length of time, the program is terminated. In the case

of a work area, this means that

- endless loops (DO, WHILE, ...),

- database accesses producing an excessively large result set,

- database accesses without a suitable index (full table scan)

do not block the processing for too long.

The system profile "rdisp/max_wprun_time" contains the maximum runtime of a

program. The

current setting is 900 seconds. Once this time limit has been exceeded,

the system tries to terminate any SQL statements that are currently

being executed and tells the ABAP processor to terminate the current

program. Then it waits for a maximum of 60 seconds. If the program is

still active, the work process is restarted.

successfully processed, the system gives it another 900 seconds.

Hence the maximum runtime of a program is at least twice the value of

the system profile parameter "rdisp/max_wprun_time".

How to correct the error

You should usually execute long-running programs as batch jobs.

If this is not possible, increase the system profile parameter

"rdisp/max_wprun_time".

Depending on the cause of the error, you may have to take one of the

following measures:

- Endless loop: Correct program;

- Dataset resulting from database access is too large:

Instead of "SELECT * ... ENDSELECT", use "SELECT * INTO internal table

(for example);

- Database has an unsuitable index: Check index generation.

You may able to find an interim solution to the problem

in the SAP note system. If you have access to the note system yourself,

use the following search criteria:

Re: Issue

Former Member — Fri, 10 Nov 2006 22:12:22 GMT

Itis using the fields as same order as interenal table

Re: Issue

Former Member — Fri, 10 Nov 2006 22:13:15 GMT

It is using fields in the same order as Internal table.But not in the same order of Bseg table

Re: Issue

Former Member — Fri, 10 Nov 2006 22:13:23 GMT

Hi,

Looks like the program is taking long time and it has exceeded the limit of 900 seconds..

If you SELECT is with in a LOOP AT ...ENDLOOP..Move it outside the loop..

I think TBKPF is an internal table..Use the FOR ALL ENTRIES ..

IF NOT TBKPF IS INITIAL.

SELECT BUKRS BELNR GJAHR BUZEI LIFNR SGTXT MWSKZ HKONT MENGE WRBTR

EBELN EBELP PRCTR WERKS SHKZG ZEKKN NPLNR KOSTL PROJK MATNR

AUFPL APLZL MEINS

FROM BSEG

INTO TABLE INTBSEG

FOR ALL ENTRIES IN TBKPF

WHERE BUKRS = TBKPF-BUKRS

AND BELNR = TBKPF-BELNR

AND GJAHR = TBKPF-GJAHR

AND MWART NE 'V'

AND TXJCD IN TXJCODE.

ENDIF.

Thanks

Naren

Re: Issue

Former Member — Fri, 10 Nov 2006 22:14:43 GMT

Are you doing this SELECT inside a LOOP? If so, would you post that code as well, please.

Rob

Re: Issue

Former Member — Fri, 10 Nov 2006 22:16:29 GMT

No this SELECT is not in the loop.

Re: Issue

Former Member — Fri, 10 Nov 2006 22:21:41 GMT

OK - would you post the section of the dump that gives the source code extract? I don't see a problem with the select.

Rob

Re: Issue

Former Member — Fri, 10 Nov 2006 22:23:54 GMT

Source code extract

004660 PERFORM TOTAL-BSET-DATA.

004670

004680 PERFORM FILL-TABLES.

004690

004700 ENDLOOP.

004710

004720 END-OF-SELECTION.

004730

004740 PERFORM MODIFY_DETAIL.

004750

004760 PERFORM PRINT-COVER-PAGE.

004770

004780 PERFORM PRINT-DETAIL-REPORT.

004790

004800 INCLUDE ZZEZ410C. "Variant capture

004810

004820 ************************************************************************

004830 * BEGIN OF FORM SUBROUTINES *

004840 ************************************************************************

004850

004860 * *

004870 * Form GET_BSEG *

004880 * *

004890 FORM GET_BSEG.

004900 * Read all record that belong to the document number

004910

004920 REFRESH: VBSEG, BSEG_NT, INTBSEG.

004930 CLEAR: VBSEG, BSEG_NT.

004940 CLEAR : QTY_TOTAL, INTBSEG_TABIX.

004950

> SELECT BUKRS BELNR GJAHR BUZEI LIFNR SGTXT MWSKZ HKONT MENGE WRBTR

004970 EBELN EBELP PRCTR WERKS SHKZG ZEKKN NPLNR KOSTL PROJK MATNR

004980 AUFPL APLZL MEINS

004990 FROM BSEG

005000 INTO TABLE INTBSEG

005010 WHERE BUKRS = TBKPF-BUKRS

005020 AND BELNR = TBKPF-BELNR

005030 AND GJAHR = TBKPF-GJAHR

005040 AND MWART NE 'V'

005050 AND TXJCD IN TXJCODE.

005060

005070 DESCRIBE TABLE INTBSEG LINES NO_OF_INTBSEG_LINES. " # 1 : Jaideep

005080

005090 LOOP AT INTBSEG.

005100 INTBSEG_TABIX = SY-TABIX. " # 2 : Jaideep Fadia 12/05/97

005110

005120 * CMS 17969 If quantity is 0, default it to 1.

005130 IF INTBSEG-MENGE = 0. "17969

005140 INTBSEG-MENGE = 1. "17969

005150 MODIFY INTBSEG INDEX SY-TABIX. "17969

Re: Issue

Former Member — Fri, 10 Nov 2006 22:27:57 GMT

OK maybe there is a problem with the select. Using NE may prevent the use of an index. Try:


SELECT  bukrs belnr gjahr buzei lifnr sgtxt mwskz hkont menge wrbtr
        ebeln ebelp prctr werks shkzg zekkn nplnr kostl projk matnr
        aufpl aplzl meins mwart
  FROM bseg
  INTO TABLE intbseg
  WHERE bukrs = tbkpf-bukrs
    AND belnr = tbkpf-belnr
    AND gjahr = tbkpf-gjahr
    AND txjcd IN txjcode.

DELETE intbseg WHERE mwart = 'V'.

You have to add MWART to the end of your internal table.

Rob

Re: Issue

Former Member — Sun, 12 Nov 2006 14:03:54 GMT

Well I don't know which release you are on and whether BSEG is still in the cluster or not.

I would try a completely different approach:

Find an Index (BSIS, BSAS, BKPF etc... ) to narrow down your extract from BSEG and give you the primary key for the BSEG records.

Retrieve all the information on the BSEG records (and I mean the full BSEG record because if the table is still clustered it will be faster) into an internal table.

Delete the records you don't need from the internal table...

Process.

Enjoy

Re: Issue

Former Member — Mon, 13 Nov 2006 18:49:39 GMT

I want change in program and see whether I can do some thing to improve perfromance of my program.

Basically some body wrote this program and initially it is SELECTING from bkpf table into internal table and then looping through that internal table and each time it is selecting from BSEG . and then from BSET.

How can I compare my performance of my program WITH and WITH OUT my changes...

my program i

Re: Issue

Former Member — Mon, 13 Nov 2006 19:08:10 GMT

Well, I asked earlier if this select was in a loop and you said no. Since this SELECT doesn't look so bad, how do you know it's not in a different SELECT. Particularly the one on BKPF. could you post that portion of the code please?

You can use the performance trace (ST05) to look at the results with and without your changes.

Rob

Re: Issue

Former Member — Mon, 13 Nov 2006 20:20:16 GMT

SELECT BUKRS BELNR GJAHR BLART BLDAT BUDAT XBLNR HWAER

FROM BKPF INTO TABLE TBKPF

WHERE BUKRS IN COMPCODE

AND BSTAT NE 'M'

AND BSTAT NE 'D'

AND BUDAT IN POSTDATE "shifted position

AND BLART IN R_BLART "5205

AND ( BLART EQ 'KG' OR "17969
BLART EQ 'KR' OR "17969
BLART EQ 'RE' ) "17969

%_HINTS ORACLE 'INDEX (BKPF BKPF______Z1)'. "17969

order by primary key.

LOOP AT TBKPF.

PERFORM GET_BSEG.

PERFORM GET_BSET.

DESCRIBE TABLE VBSEG LINES TAB_LINES.

CHECK TAB_LINES > 0.

PERFORM TOTAL-BSEG-DATA.

PERFORM TOTAL-BSET-DATA.

PERFORM FILL-TABLES.

ENDLOOP.

***************************************************************************

FORM GET_BSEG.

Read all record that belong to the document number

REFRESH: VBSEG, BSEG_NT, INTBSEG.

CLEAR: VBSEG, BSEG_NT.

CLEAR : QTY_TOTAL, INTBSEG_TABIX. " # 1 : Jaideep Fadia 11/26/97

SELECT BUKRS BELNR GJAHR BUZEI LIFNR SGTXT MWSKZ HKONT MENGE WRBTR

EBELN EBELP PRCTR WERKS SHKZG ZEKKN NPLNR KOSTL PROJK MATNR

AUFPL APLZL MEINS

FROM BSEG

INTO TABLE INTBSEG

WHERE BUKRS = TBKPF-BUKRS

AND BELNR = TBKPF-BELNR

AND GJAHR = TBKPF-GJAHR

AND MWART NE 'V'

AND TXJCD IN TXJCODE.

DESCRIBE TABLE INTBSEG LINES NO_OF_INTBSEG_LINES.

LOOP AT INTBSEG.

INTBSEG_TABIX = SY-TABIX.

CMS 17969 If quantity is 0, default it to 1.

IF INTBSEG-MENGE = 0.

INTBSEG-MENGE = 1.

MODIFY INTBSEG INDEX SY-TABIX.

ENDIF.

IF NOT INTBSEG-LIFNR IS INITIAL.

MOVE: INTBSEG-LIFNR TO TBKPF-LIFNR,

INTBSEG-SGTXT TO TBKPF-TEXT.

MODIFY TBKPF.

ELSEIF INTBSEG-MWSKZ = SPACE AND BSEG-HKONT = '0074400010'.

MOVE-CORRESPONDING INTBSEG TO BSEG_NT.

APPEND BSEG_NT.

ELSE.

RE & 0021300011

IF TBKPF-BLART = TEXT-TYP AND INTBSEG-HKONT = TEXT-GL1.

QTY_TOTAL = QTY_TOTAL + INTBSEG-MENGE.

ENDIF.

CHECK INTBSEG-MWSKZ IN TAXCODE AND INTBSEG-WRBTR NE 0."11/26

*avoid duplicate DB access AH 8/97.

IF NOT INTBSEG-EBELN IS INITIAL AND

NOT INTBSEG-EBELP IS INITIAL.

IF INTBSEG-EBELN NE EKPOTMP-EBELN AND " next record

INTBSEG-EBELP NE EKPOTMP-EBELP.

READ TABLE EKPOTMP WITH KEY EBELN = INTBSEG-EBELN

EBELP = INTBSEG-EBELP

BINARY SEARCH.

IF SY-SUBRC NE 0. " record found

EKPOTMP-EBELN = INTBSEG-EBELN.

EKPOTMP-EBELP = INTBSEG-EBELP.

CLEAR EKPO.

SELECT SINGLE WERKS INTO EKPOTMP-WERKS

FROM EKPO WHERE EBELN = INTBSEG-EBELN AND

EBELP = INTBSEG-EBELP.

IF SY-SUBRC NE 0.

CLEAR EKPOTMP-WERKS.

ENDIF.

INSERT EKPOTMP INDEX SY-TABIX.

ENDIF. " record found

ENDIF. " next record

ENDIF.

IF NOT INTBSEG-PRCTR IS INITIAL.

CHECK INTBSEG-PRCTR IN PRCENTER.

ELSEIF NOT INTBSEG-WERKS IS INITIAL.

IF INTBSEG-WERKS CO '0123456789'.

UNPACK INTBSEG-WERKS TO INTBSEG-PRCTR.

ELSE.

MOVE INTBSEG-WERKS TO INTBSEG-PRCTR.

ENDIF.

CHECK INTBSEG-PRCTR IN PRCENTER.

ELSEIF NOT EKPOTMP-WERKS IS INITIAL.

IF EKPOTMP-WERKS CO '0123456789'.

UNPACK EKPOTMP-WERKS TO INTBSEG-PRCTR.

ELSE.

MOVE EKPOTMP-WERKS TO INTBSEG-PRCTR.

ENDIF.

CHECK INTBSEG-PRCTR IN PRCENTER.

ENDIF.

CLEAR VBSEG.

IF INTBSEG-SHKZG = 'H'.

INTBSEG-WRBTR = INTBSEG-WRBTR * -1.

ENDIF.

IF INTBSEG-BUZEI EQ NO_OF_INTBSEG_LINES.

VBSEG-QTY_TOTAL = QTY_TOTAL.

ENDIF.

MOVE-CORRESPONDING INTBSEG TO VBSEG.

modify intbseg index sy-tabix.

MODIFY INTBSEG INDEX INTBSEG_TABIX." # 2

APPEND VBSEG. "apped vbseg original.

ENDIF.

ENDLOOP.

endselect.

ENDFORM. "GET_BSEG

Re: Issue

Former Member — Mon, 13 Nov 2006 20:34:47 GMT

I think your problem is here:


SELECT bukrs belnr gjahr blart bldat budat xblnr hwaer
  FROM bkpf INTO TABLE tbkpf
  WHERE bukrs IN compcode
    AND bstat NE 'M'
    AND bstat NE 'D'
    AND budat IN postdate "shifted position
    AND blart IN r_blart                                        "5205

Make sure compcode is not empty and only tests for equality.

Rather than specifying BSTAT NE ...

Specify BSTAT EQ (all the other possibilities).

Check this:

/people/rob.burbank/blog/2006/09/13/using-an-index-when-you-dont-have-all-of-the-fields

Rob

Re: Issue

Former Member — Mon, 13 Nov 2006 20:46:29 GMT

the problem is it is timing out at BSEG select .

Re: Issue

Former Member — Mon, 13 Nov 2006 21:00:51 GMT

Yes, I understand that, but it may be tha the program spent so much time on the BKPF select, it ran out of time on BSEG. So long as you looked at the issues I pointed out earlier, the BSEG SELECT looks OK.

Have you run the performance trace on this code? It should tell you where the problem is.

Rob

Re: Issue

Former Member — Mon, 13 Nov 2006 21:20:56 GMT

in st05, shall I run performance race.

Thanks.

i will take .