Sql takes long time

Question

Dear all,
The following sql takes more than 2 minutes to execute.
select * from ( select historyent0_.TRANSACTION_ID as col_0_0_, historyent0_.TRANSACTION_HISTORY_ID as col_1_0_
from TRANSACTION_HISTORY historyent0_ where (historyent0_.TRANSACTION_HISTORY_ID in
(select max(historyent1_.TRANSACTION_HISTORY_ID) from TRANSACTION_HISTORY historyent1_ group by
historyent1_.TRANSACTION_ID)) and historyent0_.TRANSACTION_STATE_ID<4 and historyent0_.USER_NAME<>:1 )
where rownum <= :2;
I have created an index
(TRANSACTION_HISTORY_ID,TRANSACTION_ID)
The Plan is :
SELECT STATEMENT  ALL_ROWSCost: 107,479  Bytes: 227.286.444  Cardinality: 5.411.582
  6 COUNT STOPKEY
    5 HASH JOIN  Cost: 107,479  Bytes: 227.286.444  Cardinality: 5.411.582
      3 VIEW VIEW SYS.VW_NSO_1 Cost: 45,87  Bytes: 76.148.033  Cardinality: 5.857.541
        2 HASH GROUP BY  Cost: 45,87  Bytes: 70.290.492  Cardinality: 5.857.541
          1 INDEX FAST FULL SCAN INDEX NIBC_FOP_AED.TEST_TRAN_HIST_IDX Cost: 8,756  Bytes: 140.464.128  Cardinality: 11.705.344
      4 TABLE ACCESS FULL TABLE NIBC_FOP_AED.TRANSACTION_HISTORY Cost: 43,995  Bytes: 156.935.878  Cardinality: 5.411.582Can someone please suggest

Answer 1

You can try this and see if it helps.
Issue is you are getting millions of rows in return and then you're doing where rownum< .... So the
query has to run in full in any case.
Try to use first rows hint if you care about getting immediate results.
Try performing fts
convert your rownum < :xyz into an actual column filter from the refined rows.....
Cheers
www.oraclefusions.com
Please visit my site for free performance tuning oracle tools.
The only real time Server-side SQL Sniffer tool developed:http://www.oraclefusions.com/applications.html#sniffer