Ah, but you haven't scrolled down to the very bottom of the page. It's quite obvious that you'll find the answer to your problems in the "Oracle Tuning, definitive reference" book (the undocumented parameter _collapse_all_updates_to_the_same_row_into_one = true, possibly). And if it's not enough (damn, the index), then I see no other solution than calling for an Oracle Health Check. SF Hemant K Chitale wrote: > > I just had a look at that code. Terrible. 6 non-indexed updates to a > custom table at logout time (when a user isn't going to wait !) > There's a better fix for the update at the end of the page -- yet the > table needs to be indexed. > > Hemant K Chitale > > http://hemantoracledba.blogspot.com > > > At 11:31 PM Friday, Niall Litchfield wrote: >> logoff triggers >> >> I've seen a database that implemented the logic shown at >> <http://www.dba-oracle.com/art_builder_sec_audit.htm>http://www.dba-oracle.com/art_builder_sec_audit.htm >> that also did not purge the audit trail. The problem with this >> particular trigger, other than that it probably doesn't do what the >> author expected or answer the requirements of SOX described at the >> start of the article, is the use of 7 non-indexed updates to the same >> large and high volume table in the logoff code one would probably >> have done. >> >> cheers >> >> Niall > > > > > > > > -- > //www.freelists.org/webpage/oracle-l > > > -- Stephane Faroult RoughSea Ltd <http://www.roughsea.com> -- //www.freelists.org/webpage/oracle-l