Performance issue on a select statement

Hi all @ SAPforums and thanks for your attention,
the task is quite simple: given a Purchase Requisition number and position (banfn, bnfpo) I have to check if a contract with the same PR as source exists in the EKPO table.
In order to check for it, I simply typed the following select:
SELECT SINGLE * FROM EKPO INTO wa_checkekpo
      WHERE bstyp EQ 'K'
        AND banfn EQ l_eban-banfn
        AND bnfpo EQ l_eban-bnfpo.
This kind of query is quite consuming (more than three seconds in my process) due to the fact that banfn and bnfpo don't belong to a key for the table.
Any idea/workaround that can lead to better performance? Please note I'm not interested in retrieving the contract number (KONNR), it's sufficient to know it exists.

Hi,
> Do not use select * if you just want to check existence of the record.
so far so good.
> Use a single variable to store teh count.
But why should we count the records if we just want to know if a key exists in
the db or not? change the second half of the recommendation to:
Use a single field of the used index for the select list to check (sy-subrc or result
of that field) if the record exists or not.
Kind regards,
Hermann

Similar Messages

Performance issue with the ABAP statements

Hello,
Please can some help me with the below statements where I am getting performance problem.
SELECT * FROM /BIC/ASALHDR0100 into Table CHDATE.
SORT CHDATE by DOC_NUMBER.
SORT SOURCE_PACKAGE by DOC_NUMBER.
LOOP AT CHDATE INTO WA_CHDATE.
   READ TABLE SOURCE_PACKAGE INTO WA_CIDATE WITH KEY DOC_NUMBER =
   WA_CHDATE-DOC_NUMBER BINARY SEARCH.
   MOVE WA_CHDATE-CREATEDON to WA_CIDATE-CREATEDON.
APPEND WA_CIDATE to CIDATE.
ENDLOOP.
I wrote an above code for the follwing requirement.
1. I have 2 tables from where i am getting the data
2.I have common fields in both the table names CREATEDON date. In both the tables I hve the values.
3. While accessing the 2 table and copying to thrid table i have to modify the field.
I am getting performance issues with the above statements.
Than
Edited by: Rob Burbank on Jul 29, 2010 10:06 AM

Hello,
try a select like the following one instead of you code.
SELECT field field2 ...
INTO TABLE it_table
FROM table1 AS T1 INNER JOIN table2 AS T2
ON t1-doc_number = t2-doc_number

Performance issue with view selection after migration from oracle to MaxDb

Hello,
After the migration from oracle to MaxDb we have serious performance issues with a lot of our tableview selections.
Does anybody know about this problem and how to solve it ??
Best regards !!!
Gert-Jan

Hello Gert-Jan,
most probably you need additional indexes to get better performance.
Using the command monitor you can identify the long running SQL statements and check the optimizer access strategy. Then you can decide which indexes might help.
If this is about an SAP system, you can find additional information about performance analysis in SAP notes 725489 and 819641.
SAP Hosting provides the so-called service 'MaxDB Migration Support' to help you in such cases. The service description can be found here:
http://www.saphosting.de/mediacenter/pdfs/solutionbriefs/MaxDB_de.pdf
http://www.saphosting.com/mediacenter/pdfs/solutionbriefs/maxDB-migration-support_en.pdf.
Best regards,
Melanie Handreck

Performance issue when using select count on large tables

Hello Experts,
I have a requirement where i need to get count of data from a database table.Later on i need to display the count in ALV format.
As per my requirement, I have to use this select count inside a nested loops.
Below is the count snippet:
LOOP at systems assigning <fs_sc_systems>.
LOOP at date assigning <fs_sc_date>.
SELECT COUNT( DISTINCT crmd_orderadm_i~header )
   FROM crmd_orderadm_i
   INNER JOIN bbp_pdigp
       ON crmd_orderadm_iclient EQ bbp_pdigpclient               "MANDT is referred as client
     AND crmd_orderadm_iguid EQ bbp_pdigpguid
     INTO w_sc_count
WHERE crmd_orderadm_i~created_at BETWEEN <fs_sc_date>-start_timestamp
     AND <fs_sc_date>-end_timestamp
     AND bbp_pdigp~zz_scsys   EQ <fs_sc_systems>-sys_name.
endloop.
endloop.
In the above code snippet,
<fs_sc_systems>-sys_name is having the system name,
<fs_sc_date>-start_timestamp is having the start date of month
and <fs_sc_date>-end_timestamp is the end date of month.
Also the data in tables crmd_orderadm_i and bbp_pdigp is very large and it increases every day.
Now,the above select query is taking a lot of time to give the count due to which i am facing performance issues.
Can any one pls help me out to optimize this code.
Thanks,
Suman

Hi Choudhary Suman ,
Try this:
SELECT crmd_orderadm_i~header
INTO it_header                 " interna table
FROM crmd_orderadm_i
INNER JOIN bbp_pdigp
    ON crmd_orderadm_iclient EQ bbp_pdigpclient
   AND crmd_orderadm_iguid   EQ bbp_pdigpguid
   FOR ALL ENTRIES IN date
WHERE crmd_orderadm_i~created_at BETWEEN date-start_timestamp
                                      AND date-end_timestamp
   AND bbp_pdigp~zz_scsys EQ date-sys_name.
    SORT it_header BY header.
    DELETE ADJACENT DUPLICATES FROM it_header
    COMPARING header.
    describe table it_header lines v_lines.
Hope this information is help to you.
Regards,
José

Increase performance of the following SELECT statement.

Hi All,
I have the following select statement which I would want to fine tune.
CHECK NOT LT_MARC IS INITIAL.
SELECT RSNUM
         RSPOS
         RSART
         MATNR
         WERKS
         BDTER
         BDMNG FROM RESB
                      INTO TABLE GT_RESB
                      FOR ALL ENTRIES IN LT_MARC
                      WHERE XLOEK EQ ' '
                      AND MATNR EQ LT_MARC-MATNR
                      AND WERKS EQ P_WERKS
                      AND BDTER IN S_PERIOD.
The following query is being run for a period of 1 year where the number of records returned will be approx 3 million. When the program is run in background the execution time is around 76 hours. When I run the same program dividing the selection period into smaller parts I am able to execute the same in about an hour.
After a previous posting I had changed the select statement to
CHECK NOT LT_MARC IS INITIAL.
SELECT RSNUM
         RSPOS
         RSART
         MATNR
         WERKS
         BDTER
         BDMNG FROM RESB
                      APPENDING TABLE GT_RESB PACKAGE SIZE LV_SIZE
                      FOR ALL ENTRIES IN LT_MARC
                      WHERE XLOEK EQ ' '
                      AND MATNR EQ LT_MARC-MATNR
                      AND WERKS EQ P_WERKS
                      AND BDTER IN S_PERIOD.
ENDSELECT.
But the performance improvement is very negligible.
Please suggest.
Regards,
Karthik

Hi Karthik,
<b>Do not use the appending statement</b>
Also you said if you reduce period then you get it quickly.
Why not try dividing your internal table LT_MARC into small internal tables having max 1000 entries.
You can read from index 1 - 1000 for first table. Use that in the select query and append the results
Then you can refresh that table and read table LT_MARC from 1001-2000 into the same table and then again execute the same query.
I know this sounds strange but you can bargain for better performance by increasing data base hits in this case.
Try this and let me know.
Regards
Nishant
> I have the following select statement which I would
> want to fine tune.
>
>   CHECK NOT LT_MARC IS INITIAL.
> SELECT RSNUM
>          RSPOS
> RSART
>          MATNR
> WERKS
>          BDTER
> BDMNG FROM RESB
>                       INTO TABLE GT_RESB
> FOR ALL ENTRIES IN LT_MARC
>                       WHERE XLOEK EQ ' '
> AND MATNR EQ LT_MARC-MATNR
>                       AND WERKS EQ P_WERKS
> AND BDTER IN S_PERIOD.
>
> e following query is being run for a period of 1 year
> where the number of records returned will be approx 3
> million. When the program is run in background the
> execution time is around 76 hours. When I run the
> same program dividing the selection period into
> smaller parts I am able to execute the same in about
> an hour.
>
> After a previous posting I had changed the select
> statement to
>
>   CHECK NOT LT_MARC IS INITIAL.
> SELECT RSNUM
>          RSPOS
> RSART
>          MATNR
> WERKS
>          BDTER
> BDMNG FROM RESB
> APPENDING TABLE GT_RESB
>   PACKAGE SIZE LV_SIZE
>                     FOR ALL ENTRIES IN LT_MARC
>   WHERE XLOEK EQ ' '
>                     AND MATNR EQ LT_MARC-MATNR
>   AND WERKS EQ P_WERKS
>                     AND BDTER IN S_PERIOD.
> the performance improvement is very negligible.
> Please suggest.
>
> Regards,
> Karthik
Hi Karthik,

Performance issue with the Select query

Hi,
I have an issue with the performance with a seclet query.
In table AFRU - AUFNR is not a key field.
So i had selected the low and high values into s_reuck and used it in Where condition.
Still i have an issue with the Performance.
SELECT SINGLE RUECK
RMZHL
IEDD
AUFNR
STOKZ
STZHL
FROM AFRU INTO table t_AFRU
FOR ALL ENTRIES IN T_ZSCPRT100
WHERE RUECK IN S_RUECK AND
AUFNR = T_ZSCPRT100-AUFNR AND
STOKZ = SPACE AND
STZHL = 0.
I had also cheked by createing an index for AUFNR in the table AFRU...it does not help.
Is there anyway that we can declare Key field while declaring the Internal table....?
ANy suggestions to fix the performance issue is apprecaited!
Regards,
Kittu

Hi,
Thank you for your quick response!
Rui dantas, i have lill confusion...this is my code below :
data : t_zscprt type standard table of ty_zscprt,
       wa_zscprt type ty_zscprt.
types : BEGIN OF ty_zscprt100,
        aufnr type zscprt100-aufnr,
        posnr type zscprt100-posnr,
        ezclose type zscprt100-ezclose,
        serialnr type zscprt100-serialnr,
        lgort type zscprt100-lgort,
      END OF ty_zscprt100.
data : t_zscprt100 type standard table of ty_zscprt100,
       wa_zscprt100 type ty_zscprt100.
Types: begin of ty_afru,
            rueck type CO_RUECK,
            rmzhl type CO_RMZHL,
            iedd type RU_IEDD,
            aufnr type AUFNR,
            stokz type CO_STOKZ,
            stzhl type CO_STZHL,
         end of ty_afru.
data : t_afru type STANDARD TABLE OF ty_afru,
        WA_AFRU TYPE TY_AFRU.
SELECT AUFNR
        POSNR
        EZCLOSE
        SERIALNR
        LGORT
        FROM ZSCPRT100 INTO TABLE T_ZSCPRT100
        FOR ALL ENTRIES IN T_ZSCPRT
        WHERE   AUFNR = T_ZSCPRT-PRTNUM
        AND   SERIALNR IN S_SERIAL
        AND    LGORT   IN S_LGORT.
    IF sy-subrc <> 0.
       MESSAGE ID 'Z2' TYPE 'I' NUMBER '41'. "BDCG87
       stop."BDCG87
    ENDIF.
ENDIF.
SELECT    RUECK
              RMZHL
              IEDD
              AUFNR
              STOKZ
              STZHL
              FROM AFRU INTO TABLE T_AFRU
              FOR ALL ENTRIES IN T_ZSCPRT100
              WHERE RUECK IN S_RUECK     AND
                    AUFNR = T_ZSCPRT100-AUFNR AND
                    STOKZ = SPACE AND
                    STZHL = 0.
Using AUFNR, get AUFPL from AFKO
Using AUFPL, get RUECK from AFVC
Using RUEKC, read AFRU
In other words, one select joining AFKO <-> AFVC <-> AFRU should get what you want.
This is my select query, would you want me to write another select query to meet this criteria..
From AUFNR> I will get AUFPL from AFKO> BAsed on AUFPL I will get RUECK, based on RUEKC i need to read AFRU..but i need to select few field from AFRu based on AUFNR....
ANy suggestions wil be appreciated!
Regards
Kittu

TDE Issue with UPDATE/SELECT statement

We just implemented TDE on a table and now our import script is getting errors. The import script has not changed and has been running fine for over a year. The script failed right after applying TDE on the table.
Oracle 10g Release 2 on Solaris.
Here are the encrypted colums:
COLUMN_NAME ENCRYPTION_ALG SALT
PERSON_ID AES 192 bits key NO
PERSON_KEY AES 192 bits key NO
USERNAME AES 192 bits key NO
FIRST_NAME AES 192 bits key NO
MIDDLE_NAME AES 192 bits key NO
LAST_NAME AES 192 bits key NO
NICKNAME AES 192 bits key NO
EMAIL_ADDRESS AES 192 bits key NO
AKO_EMAIL AES 192 bits key NO
CREATION_DATE AES 192 bits key NO
Here is the UPDATE/SELECT statement that is failing:
UPDATE cslmo_framework.users a
       SET ( person_id
           , username
           , first_name
           , middle_name
           , last_name
           , suffix
           , user_status_seq
         = (
             SELECT person_id
                  , username
                  , first_name
                  , middle_name
                  , last_name
                  , suffix
                  , user_status_seq
               FROM cslmo.vw_import_employee i
              WHERE i.person_key = a.person_key
     WHERE EXISTS
               SELECT 1
                 FROM cslmo.vw_import_employee i
                WHERE i.person_key = a.person_key
                  AND (    NVL(a.person_id,0)        <> NVL(i.person_id,0)
                        OR NVL(a.username,' ')       <> NVL(i.username,' ')
                        OR NVL(a.first_name,' ')     <> NVL(i.first_name,' ')
                        OR NVL(a.middle_name,' ')    <> NVL(i.middle_name,' ')
                        OR NVL(a.last_name,' ')      <> NVL(i.last_name,' ')
                        OR NVL(a.suffix,' ')         <> NVL(i.suffix,' ')
                        OR NVL(a.user_status_seq,99) <> NVL(i.user_status_seq,99)
cslmo@awpswebj-dev> exec cslmo.pkg_acpers_import.p_users
Error importing USERS table.START p_users UPDATE
Error Message: ORA-01483: invalid length for DATE or NUMBER bind variableI rewrote the procedure using BULK COLLECT and a FORALL statement and that seems to work fine. Here is the new code:
declare
   bulk_errors EXCEPTION ;
   PRAGMA EXCEPTION_INIT(bulk_errors,-24381) ;
   l_idx      NUMBER ;
   l_err_msg VARCHAR2(2000) ;
   l_err_code NUMBER ;
   l_update   NUMBER := 0 ;
   l_count    NUMBER := 0 ;
   TYPE person_key_tt
       IS
           TABLE OF cslmo_framework.users.person_key%TYPE
                INDEX BY BINARY_INTEGER ;
   arr_person_key   person_key_tt ;
   TYPE person_id_tt
       IS
          TABLE OF cslmo_framework.users.person_id%TYPE
                INDEX BY BINARY_INTEGER ;
   arr_person_id   person_id_tt ;
   TYPE username_tt
      IS
          TABLE OF cslmo_framework.users.username%TYPE
               INDEX BY BINARY_INTEGER ;
   arr_username   username_tt ;
   TYPE first_name_tt
      IS
         TABLE OF cslmo_framework.users.first_name%TYPE
              INDEX BY BINARY_INTEGER ;
   arr_first_name   first_name_tt ;
   TYPE middle_name_tt
     IS
         TABLE OF cslmo_framework.users.middle_name%TYPE
             INDEX BY BINARY_INTEGER ;
   arr_middle_name   middle_name_tt ;
   TYPE last_name_tt
         IS
            TABLE OF cslmo_framework.users.last_name%TYPE
                 INDEX BY BINARY_INTEGER ;
   arr_last_name   last_name_tt ;
   TYPE suffix_tt
         IS
            TABLE OF cslmo_framework.users.suffix%TYPE
                 INDEX BY BINARY_INTEGER ;
   arr_suffix   suffix_tt ;
   TYPE user_status_seq_tt
         IS
            TABLE OF cslmo_framework.users.user_status_seq%TYPE
                 INDEX BY BINARY_INTEGER ;
   arr_user_status_seq   user_status_seq_tt ;
   CURSOR users_upd IS
      SELECT i.person_key
             ,i.person_id
             ,i.username
             ,i.first_name
             ,i.middle_name
             ,i.last_name
             ,i.suffix
             ,i.user_status_seq
      FROM   cslmo.vw_import_employee i ,
             cslmo_framework.users    u
      WHERE i.person_key = u.person_key ;
begin
   OPEN users_upd ;
   LOOP
        FETCH   users_upd
         BULK
      COLLECT
         INTO    arr_person_key
               , arr_person_id
               , arr_username
               , arr_first_name
               , arr_middle_name
               , arr_last_name
               , arr_suffix
               , arr_user_status_seq
        LIMIT         100 ;
        FORALL idx IN 1 .. arr_person_key.COUNT
            SAVE EXCEPTIONS
            UPDATE cslmo_framework.users u
              SET
                   person_id                =   arr_person_id(idx)
                 , username                 =   arr_username(idx)
                 , first_name               =   arr_first_name(idx)
                 , middle_name              =   arr_middle_name(idx)
                 , last_name                =   arr_last_name(idx)
                 , suffix                   =   arr_suffix(idx)
                 , user_status_seq          =   arr_user_status_seq(idx)
             WHERE u.person_key = arr_person_key(idx)
             AND
                   ( NVL(u.person_id,0) != NVL(arr_person_id(idx),0)
             OR
                     NVL(u.username,' ') != NVL(arr_username(idx),' ')
             OR
                     NVL(u.first_name,' ') != NVL(arr_first_name(idx),' ')
             OR
                     NVL(u.middle_name, ' ') != NVL(arr_middle_name(idx), ' ')
             OR
                     NVL(u.last_name,' ') != NVL(arr_last_name(idx),' ')
             OR
                     NVL(u.suffix,' ') != NVL(arr_suffix(idx),' ')
             OR
                     NVL(u.user_status_seq,99) != NVL(arr_user_status_seq(idx),99)
      l_count := arr_person_key.COUNT ;
      l_update := l_update + l_count ;
      EXIT WHEN users_upd%NOTFOUND ;
   END LOOP ;
   CLOSE users_upd ;
   COMMIT ;
   dbms_output.put_line('updated records: ' || l_update);
   EXCEPTION
      WHEN bulk_errors THEN
           FOR i IN 1 .. sql%BULK_EXCEPTIONS.COUNT
           LOOP
              l_err_code   :=   sql%BULK_EXCEPTIONS(i).error_code ;
              l_err_msg    :=   sqlerrm(-l_err_code) ;
              l_idx        :=   sql%BULK_EXCEPTIONS(i).error_index;
              dbms_output.put_line('error code: ' || l_err_code);
              dbms_output.put_line('error msg: ' || l_err_msg);
              dbms_output.put_line('at index: ' || l_idx);
           END LOOP ;
           ROLLBACK;
           RAISE;
end ;
cslmo@awpswebj-dev> @cslmo_users_update
updated records: 1274There are about 20 or so other procedure in the import script. I don't want to rewrite them.
Does anyone know why the UPDATE/SELECT is failing? I checked Metalink and could not find anything about this problem.

This is now an Oracle bug, #9182070 on Metalink.
TDE (transparent data encryption) does not work when an update/select statement references a remote database.

Performance problem of asset selection statement.

Hi guys quick question.
SELECT a~bukrs
           a~anln1
           b~anln2
           a~anlkl
           a~aktiv
           a~txt50
           a~lvtnr
           b~afabe
           b~afabg
           b~ndjar
           b~ndper
           INTO TABLE it_anla
           FROM anla AS a
           INNER JOIN anlb AS b
           ON a~bukrs EQ b~bukrs
           AND a~anln1 EQ b~anln1
           WHERE a~bukrs IN s_bukrs
           AND a~anln1 IN s_anln1
           AND a~anln2 EQ '0000'
           AND a~anlkl IN s_anlkl
           AND a~aktiv EQ '00000000'
           AND a~lvtnr IN s_lvtnr
           AND b~afabe EQ 01.
I have a select statement which is filtered by capitalization date which is the field AKTIV. This select alone has 450 thousand hits. Having an ANLN2 = 0 and AKTIV = blank/no date means that the asset group is no longer active.
Now for another scenario I have to retrieve table ANLA again excluding all the Main Asset Number/ANLN1 found on that table.
Is there a way to select by doing it only once?
I tried to pass all the anln1 to a range table but the program dumps. I think the range table can't handle too many entries.
Retrieving all the entries from the DB then processing it takes longer.
I tried to delete the table using a loop but it takes to long as it process the table every loop.
LOOP AT it_anla WHERE anln2 EQ '0000' AND aktiv EQ '0000000'.
      DELETE it_anla WHERE bukrs EQ it_anla-bukrs AND anln1 EQ it_anla-anln1.
    ENDLOOP.
Thanks.
Edited by: Thomas Zloch on Sep 21, 2010 5:39 PM - please use code tags

Moderator message - Welcome to SCN
If the range table for anln1 is large and contains distinct values, you can try using it in a FOR ALL ENTRIES construct rather than IN.
Rob

Query related to excution and performance of count in select statement

Hi,
What are the difference between
select count(1) from table
and
select count(*) from table

Thanks for the reply.I searched lots of thread for the same and found that it's same (excution wise and performance wise).
But in few thread i found that someone is saying
"At the time of exceution Count(1) will conevrted internally to count(*) " --- is it correct
"is there any soecial significance of using count(42)" ?
Tx,
Anit

Performance issue with Printing AR Statements

We are using XML Publisher 5.2. We are usign XML Publisher to print our AR statements.
To print our statements, it took us nearly 6-8 hrs for the entire process.
Is there any performance tuning parameters, that we should set in order to improve the performance.
Thanks,
- vasu -

Thats way too long ... should have been 1/2 hr at most based on our tests. Please log a TAR and be sure to load template and the XML data (yes all of it :o) and we can take a look.
Regards, Tim

Performance issue doing a prepare statement with LIKE

Hi
I'm doing a preparestatement like this
select foo
from bar
where des like ?
and then do the prepare with the string 'foo%'
If I do the prepare of this variable, the query is very slow. If I don't do the prepare, the query is very fast. Can anyone help me out on this? DBA's will kill me if I don't use prepare statements but users will kill be if the query is slow...
Thank you!
HappyGuy

As indicated in the last reply this probably has to do with use of string literals over bound variables
With a prepared statement the optimiser dosen't get the chance to use histograms against your data... so whilst it might be more efficent to drive off an index for some of the searches it looses access to the data (histograms) that enable it to make this decission...
Hints are probably the way to go... if your sure that driving off a index is always going to be the most efficent...
Using explain (inside of JDeveloper9i or turning on the autotrace functionality inside of sqlplus) will give you some idea of how Oracle would drive the query if you use a string literal
Dom

Performance Issue: Select From BSEG & BKPF

Hi experts,
Performance issue on the select statements; how can I improve the performance?
Select Company Code (BUKRS)
Accounting Document Number (BELNR)
                    Document Type (BLART)
                    Posting Date in the Document (BUDAT)
                    Document Status (BSTAT)
                    Reversal Document or Reversed Document Indicator (XREVERSAL)
           From Accounting Document Header (BKPF)
             Into I_BKPF
         Where BKPF-BUKRS = I_VBAK-BUKRS_VF
       BKPF-BLART = KI
                    BKPF-BUDAT = SY-DATUM 2 days
                    BKPF-BSTAT = Initial
                    BKPF-XREVERSAL <> 1 or 2
Select Company Code (BUKRS)
                    Accounting Document Number (BELNR)
                    Assignment Number (ZUONR)
                    Sales Document (VBEL2)
                    Sales Document Item (POSN2)
                    P & L Statement Account Type (GVTYP)
            From Accounting Document Segment (BSEG)
              Into I_BSEG
          Where BSEG-BUKRS = I_VBAK-BUKRS
        BSEG-VBELN = I_VBAK-VBEL2
        BSEG-POSN2 = I_VBAP-POSNR
             BSEG-BELNR = I_BKPF-BELNR
                     P & L Statement Account Type (GVTYP) = X

Hi,
to improve the performance, you can use the secondary indices viz., BSIK / BSAK, BSID / BSAD, BSIS.
Hope this helps.
Best Regards, Murugesh AS

Multiple table select Performance Issue

Hi,
I would like to get an opinion which from these query which is faster and has a performance issue..
SELECT EMP_ID, NAME, DEPT_NAME
FROM EMP, DEPT
WHERE EMP_ID = DEPT_ID;
or
SELECT EMP_ID, NAME, (SELECT DEPT_NAME FROM DEPT WHERE ID = P_ID)
FROM EMP
WHERE EMP_ID = P_ID;

lets say that EMP_ID on Dept table is linked to EMP_ID table on EMP..Well...I don't get your design, but the two queries may return different results.
Comparing the performance doesn't make sense.
Nevertheless, the only way is to run them both and see which one is faster or see which one has the lowest IO.
There's no way we can tell you which is faster by just looking at the text of the queries.
Post some explain plans or traces.

Secondary Index Select Statement Problem

Hi friends.
I have a issue with a select statement using secondary index,
SELECT SINGLE * FROM VEKP WHERE VEGR4 EQ STAGE_DOCK
                                  AND VEGR5 NE SPACE
                                  AND WERKS EQ PLANT
        %_HINTS ORACLE
        'INDEX("&TABLE&" "VEKP~Z3" "VEKP^Z3" "VEKP_____Z3")'.
given above statement is taking long time for processing.
when i check for the same secondary index in vekp table i couldn't see any DB index name with vekp~z3 or vekp^z3 or vekp____z3.
And the sy-subrc value after select statement is 4. (even though values avaliable in VEKP with given where condition values)
My question is why my select statement is taking long time and sy-subrc is 4?
what happens if a secnodary index given in select statement, which is not avaliable in that DB Table?

Hi,
> ONe more question: is it possible to give more than one index name in select statement.
yes you can:
read the documentation:
http://download.oracle.com/docs/cd/A97630_01/server.920/a96533/hintsref.htm#5156
index_hint:
This hint can optionally specify one or more indexes:
- If this hint specifies a single available index, then the optimizer performs
a scan on this index. The optimizer does not consider a full table scan or
a scan on another index on the table.
- If this hint specifies a list of available indexes, then the optimizer
considers the cost of a scan on each index in the list and then performs
the index scan with the lowest cost. The optimizer can also choose to
scan multiple indexes from this list and merge the results, if such an
access path has the lowest cost. The optimizer does not consider a full
table scan or a scan on an index not listed in the hint.
- If this hint specifies no indexes, then the optimizer considers the
cost of a scan on each available index on the table and then performs
the index scan with the lowest cost. The optimizer can also choose to
scan multiple indexes and merge the results, if such an access path
has the lowest cost. The optimizer does not consider a full table scan.
Kind regards,
Hermann

How to solve performance issue

Hi
        How to resolve performance issue in following select query-
SELECT *
          INTO CORRESPONDING FIELDS OF TABLE it_final
          FROM ce1zcsc
          WHERE paledger EQ c_10                 "Currency Type
          AND   vrgar     IN s_vrgar              "Record Type
          AND   versi     EQ space                "Plan Version
          AND   perio     IN r_perio              "Period
          AND   bukrs     IN s_bukrs.             "Company Code
TABLE CE1ZCSC has around 173 fields,but it_final has around 105 fields.
The indexes are created for the following fields:
paledger
vrgar
versi
perio
bukrs and
prctr.
I doubt whether we should look for Estim. CPU-Costs in index range scan or table access by index rowid in Execution plan for SQL statement.
If anybody can provide me with informative documents on performance issue.

Hi,
Dont use " * " & " corresponding fields " in the select query rather declare all the fields and use the "into table" clause.
Let me know if you still face the same problem.
-Naveen.

Performance issue on a select statement

Similar Messages

Maybe you are looking for