On 08/15/2004 12:48:21 PM, Wolfgang Breitling wrote: > Mladen, > > Can you explain a bit what your example is intended to show in response to > my post. As far as I can tell it merely confirms what I had been saying > a) dynamic_sampling=4 does not necessarily guarantee that analyzed tables > will be sampled. > Both tables in your example have no statistics, so they DO get sampled > once you force the CBO Wolfgang, my impression was that your original statement was that dynamic sampling doesn't work. I just wanted to show that it does work, provided the use of CBO is forced. Unless the CBO is conscienceously forced, the database will use RBO if none of the tables have statistics. I'm experimenting with ALL_ROWS and and OPTIMIZER_INDEX parameters to get the desired database behavior. My ultimate goal is to convert an 8i database using RBO (forced directly by parameter and a periodic job that deletes statistics from any new table) into a 10g database using CBO, without much degradation in query performance. -- Mladen Gogala Oracle DBA ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------