Trace Evaluation

2001-06-25 Thread VIVEK_SHARMA
Is there Any Scope for Improvement in the Following ? select acct_crncy_code into :b0 from GAM where acid=:b1 call count cpuelapsed disk querycurrent rows --- -- -- -- -- -- -- Parse1 0.00

RE: Trace Evaluation

2001-06-25 Thread Shevtsov, Eduard
Hi, Do you think you could decrease number of executions ? 4 blocks per fetch => probably index's blevel = 2 Is partitioning available for you? You may partition the table on acid column (I guess it's primary key). So you get local partitioned index on acid with smaller partitions Regards, Ed

RE: Trace Evaluation

2001-06-25 Thread Koivu, Lisa
Title: RE: Trace Evaluation Hi Vivek, Yes.  Select into expects one record (and your index is unique), so you can do this: Try adding AND ROWNUM < 2 to the query and it will quit after it finds the record it is looking for.  Please post your results, I'm curious how big of a di

RE: Trace Evaluation

2001-06-26 Thread VIVEK_SHARMA
ents of list ORACLE-L > Subject: RE: Trace Evaluation > > Hi, > > Do you think you could decrease number of executions ? > > 4 blocks per fetch => probably index's blevel = 2 > > Is partitioning available for you? You may partition the table on > acid col

RE: Trace Evaluation

2001-06-26 Thread Shevtsov, Eduard
Hi Vivek, you have the best plan that you can get via index path. I'm afraid, there are only 2 way here that I can offer for you. 1. Decrease number of executions. Is such huge number of executions justified in your case? 2. If you can't change application logic (don't have access to SQL code), y