Re: UNUSABLE Indexes

From: John K. Hinsdale <>
Date: Thu, 27 Dec 2007 11:17:37 -0800 (PST)
Message-ID: <>

On Dec 27, 12:01 pm, "Jonathan Lewis" <> wrote:
> the update to the index after
> the insert /*+ append */ is a special ...
> by sorting the new entries and using a bulk-insert
> into the index leaf blocks).
> If the table has been truncated then the index has also been
> truncated, and the optimised update is similar in resource
> usage to a 'create' - however, create index does not generate
> undo, so will be a little more efficient. Moreover, the optimized
> update has to be logged, but if you end up with  a create
> (or rebuild)  you can choose to do it nologging.

OK, thanks that was very clear & enlightening.

      --JH Received on Thu Dec 27 2007 - 13:17:37 CST

Original text of this message