On Fri, Jul 18, 2003 at 08:26:59AM -0300, Vilson farias wrote:
> Greetings,
>
> I've been fighting against a very strange behaviour found in PostgreSQL
> 7.1.2 on a RedHat 6.2. I have a very simple table called site_site and I
> lost it's indexes everytime I run a vaccum. Do you know why this happens? Is
> there a way to get around or fix this kind of problem?I put a full sequence
> of steps I've used to make it happen.
The indexes are not lost; after you have done an ANALYZE, the system
knows that they won't speed up your queries, so it chooses sequential
scans instead. If you put a lot of data in the table and ANALYZE again,
you'll find that the indexes are used.
Richard
