Hello, Database version is 11g2. I tried histograms. But the problem here again is that columns are kind-of intercorrelated. For example, I have 1 column named "country", then other columns have different stats for each country. So in fact I need count(distinct country) statistic sets for proper estimation. I tried partitioning by country, but it has drawbacks in my case which slow down at other places. Best regards Ilya 2010/5/12 Marcin Przepiorowski <pioro1@xxxxxxxxx>: > On Wed, May 12, 2010 at 8:53 PM, Ilia Kantor <iliakan@xxxxxxxxx> wrote: >> Hello, >> >> I have a multi-predicate query like: >> >> SELECT * FROM table WHERE >> a BETWEEN 5 and 10 >> and >> b in (3,4,5) >> and c in (9) >> ... >> >> Unfortunately, optimizer is not very good at estimating cardinalities. >> Columns are somewhat intercorrelated and "multiple column stats" >> feature does not fix that. >> >> Although, I gather my own stats which give me exact idea for >> selectivity of each predicate. >> How can I pass this information to optimizer? >> > > Hi, > > What database version ? > Did you try change gathering statistic from estimating to calculating > and use histograms ? > > > -- > Marcin Przepiorowski > http://oracleprof.blogspot.com > -- //www.freelists.org/webpage/oracle-l