Re: Partial Analyze bad?

From: David Lane <dlane_at_cix.compulink.co.uk>
Date: Tue, 2 Mar 1999 19:43:07 GMT
Message-ID: <F7zG3v.79I@cix.compulink.co.uk>

In article <36D6237E.E6DEFE13_at_cadvision.com>, krachyn_at_cadvision.com (Ken Rachynski) wrote:

> Good day,
>
> I was trying to analyze the tables in my application the other day and
> had to stop the process with only half of the tables analyzed. Since
> then, nearly all operations that use a mixture of analyzed and
> non-analyzed tables slowed down considerably. I have removed all
> statistics on my tables and the performance seems to have improved.
>
> Would I benefit much by having all the tables analyzed? Is it bad to
> have only some analyzed?
>
> --
> Ken Rachynski
> Database Analyst
> krachyn_at_cadvision.com
> <http://www.cadvision.com/krachyn>
> ICQ: 3113514
>
>

The optimiser definitely gets confused when only some of the tables are analysed .... it's worth analysing all of your tables at the beginning, then you can do so many each day to keep things up-to-date

Dave Lane (dlane_at_pt.lu) Received on Tue Mar 02 1999 - 13:43:07 CST