Table Replicated,index not used-BitOld

Hello Team

I need your help on this. We were on 9.2.0.1 on solaris and had two instances running on two separate machines; One was production (PRD) and the other for administration/accounting related (SHD). We have table replicated from PRD to SHD; The number of distinct keys in both were the same and so is the number of rows ; The query on prd returns results in few seconds while on the shd instance it took long time without any hints ; with hint it processes faster. The clustering factor was high

[1]http://www.geocities.com/dba_assist/oracle-l/prd.html

[2]http://www.geocities.com/dba_assist/oracle-l/shd.html

First i rebuilt the index ; There was no improvement in the statistics. Dropped , Recreated, Analyzed .. No improvement. We later migrated to 9.2.0.6 and now both the instances are taking same(=more) time and doing FTS. If the optimizer could use the index on the first occassion in PRD+9201 why it does not do it now?

Any suggestions would be appreciated.

TIA GovindanK

References

1. http://www.geocities.com/dba_assist/oracle-l/prd.html
2. http://www.geocities.com/dba_assist/oracle-l/shd.html

--
http://www.freelists.org/webpage/oracle-l