dbms_stats vs analyze

From: cj <oraquest_c_at_yahoo.com>
Date: 13 Feb 2005 09:20:39 -0800
Message-ID: <1108315239.710244.89160@c13g2000cwb.googlegroups.com>

Hello,

DBMS - 9.2.0.6 HP/UX Super Dome, Data Warehouse. I have a difference in counts in several of our partitioned tables. If I query the dba_tab_partitions and sum all the row counts, count is over 15 mil. But if I query dba_tables I only get 55,000 rows. The analyze script is set to analyze the tables estimating a sample of 5% this runs 2x's a week. But the partitions are analyzed via dbms_stats only after the partition has been fully populated or the month is completed. If I do a row count from these tables it's pretty close to what the dba_tab_partitions sum equates to.

Why is there such a difference in the amounts? Any recommendations? Are we going about this the wrong way?
CJ Received on Sun Feb 13 2005 - 11:20:39 CST