That is exactly the set of circumstances that REVERSE key indexes are designed for. Go for it, brother! on 8/26/05 5:49 AM, Hemant K Chitale at hkchital@xxxxxxxxxxxxxx wrote: > > I have frequently wondered if REQUEST_ID in the Oracle Applications table > FND_CONCURRENT_REQUESTS > would be a good candidate to index using a Reverse Key index. There are > certain peak times in my > organisation's ERP processing when 3500 requests may be submitted within > the space of 30 to 50 minutes > {Event Alert Triggers}. > This column is a Unique Key (could actually have been defined as a Primary > Key) and is queried by single > values, not by range scans. {Other queries on the table may be executing > Range Scans on the > REQUESTED_START_DATE or ACTUAL_START_DATE (for daily "Performance of > Concurrent Requests" reports) > or USER_ID columns} > > Should I try rebuilding FND_CONCURRENT_REQUESTS_U1 on REQUEST_ID as a > Reverse Key Index ? > Opinions ? > > Hemant -- //www.freelists.org/webpage/oracle-l