On 02/11/05, Steve Adams <steve.adams@xxxxxxxxxxxx> wrote: > > We got 2 redo records for the BEFORE trigger (but only 1 for the AFTER > trigger). > An extra redo record was needed because the row was locked separately before > the > application of the change vectors for the update. > I hadn't been aware of that, I'll try to remember for the future. Thanks I take it you're not arguing against the need for a trigger to enforce the data integrity of calculated fields? Stephen -- It's better to ask a silly question than to make a silly assumption. -- //www.freelists.org/webpage/oracle-l