I am trying to implement histograms and have been following Burleson's method of implementing. I'm on Oracle 9i Rel 2 Ver 188.8.131.52 on Windows 2000 Advanced Server. I have run :
ownname => 'SNAP',
estimate_percent => dbms_stats.auto_sample_size,
method_opt => 'for all columns size skewonly',
degree => 4
And am trying to interpret what I am seeing in DBA_histograms. I am the first one to run histograms on this db and there are only about 300 tables in this schema with about 15 columns or so per table. I am ending up with a lot of rows in this table. I wasn't expecting that. I am expecting one row per histogram and one histogram per column. Is that right? Can histograms be deleted? Using delete schema stats didn't do it....