From: "Jonathan Lewis" Newsgroups: comp.databases.oracle.server Subject: Re: On Indexes and Commits Date: Wed, 8 Aug 2001 14:48:12 +0100 Message-ID: <997278489.28573.0.nnrp-13.9e984b29@news.demon.co.uk> References: NNTP-Posting-Host: jlcomp.demon.co.uk X-NNTP-Posting-Host: jlcomp.demon.co.uk:158.152.75.41 X-Trace: news.demon.co.uk 997278489 nnrp-13:28573 NO-IDENT jlcomp.demon.co.uk:158.152.75.41 X-Complaints-To: abuse@demon.net X-Newsreader: Microsoft Outlook Express 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Lines: 39 General principle. when you insert a row into a table, you insert the index entries at the same time. There are minor variations on a theme, but as each INSERT statement completes, the table and its indexes are self-consistent. Rolling back is very expensive - start looking into TEMPORARY TABLES with 'on commit delete rows'. -- Jonathan Lewis Seminars on getting the best out of Oracle Last few places available for Sept 10th/11th See http://www.jlcomp.demon.co.uk/seminar.html Kirt Thomas wrote in message ... >We're sitting here puzzling over this :) > >Our developers use a table to store data during the execution of a >report, they may end up inserting several to tens of thousands of >records into this table to be reported on (of course this is a bad, >but that's neither here nor there right now). The way the dev's are >handling the records in this 'temp' table is to do no commit, and >issue a rollback after the report is done - effectively cancelling the >inserts. This allows multiple reports to run at the same time as each >report only see's it's own sessions data. > >My assertion is that doing things this effectively negates any indexes >on the table. I don't see how they could be used w/out commiting the >data. Is this true? Thanks.