Rather than guessing, you can know the answer. Just run 10053 tracing in both instances and compare. To interpret the results, I highly recommend a paper written by Wolfgang Breitling, available at http://www.centrexcc.com/A%20Look%20under%20the%20Hood%20of%20CBO%20-%20the%2010053%20Event.pdf Ron Crisco On Tue, Apr 13, 2010 at 9:54 AM, Joan Hsieh <joan.hsieh@xxxxxxxxx> wrote: > Hi dear listers, > > I have a very strange performance issue which I couldn't figure it out. > I have 2 11.2.0.1 databases with the exactly same parameters, they both on > the same server. One sql statement run on both database generated 2 diff > plans. So I exported the whole schema stats from the good db, imported it to > the bad one. It didn't change the result. I can easily fix the performance > issue on the bad db by setting optimizer_feature_enabled to 10g, but I just > wondering why the 2 databases have different plans even after I imported the > stats and checked the stats. > > Thanks, > > Joan > > > > -- > //www.freelists.org/webpage/oracle-l > > >