Thanks, Connor and Wolfgang! I did not know this. My tests look like Connor's (in my 9.2..0.4 database) If a table does not have stats, some statistical information is updated for that table if you compute statistics on an index.=20 num_rows and blocks appear to be correct. avg_row_len is set to 100.=20 Some other columns look like they're zero-filled (empty_blocks, avg_space, chain_cnt, avg_space_freelist_blocks, num_freelist_blocks.) If the table DOES have stats, computing statistics on the index does NOT overwrite the statistics (at least in my tests). It does, however, update the last_analyzed column. Guess I need to be more careful of the order of my steps when I move tables and indexes!! Barb On 6/8/05, Connor McDonald <mcdonald.connor@xxxxxxxxx> wrote: > I remember on 8.1.x using "comp stats" on an index rebuild also did > the table (unless that table was partitioned...) >=20 >=20 > (on an unrelated note - how do we post in monospaced font ... does <pre> = wo=3D > rk ?) >=20 > hth > connor=3D20 >=20 > On 6/9/05, Barbara Baker <barb.baker@xxxxxxxxx> wrote: > > Ok, I'll start running some tests. -- //www.freelists.org/webpage/oracle-l