"Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message news:<pan.2003.02.20.19.11.31.525960_at_yahoo.com.au>...
> On Thu, 20 Feb 2003 16:25:55 +0100, Jan Gelbrich wrote:
>
> > Hi, I would like to know not only how often, but which schemas should be
> > analyzed ...
> >
> > .. it is because I heard a couple of times that
> > analyzing SYSTEM
> > can bring performance down,
> > so that You should drop those statistics.
> >
> > My question is:
> >
> > 1. *Why* can analyze of SYSTEM be a reason at all to bring performance down
> > ?
>
> In general, it is an extremely bad idea to take code which has been
> written and tuned with the RULE-based optimizer in mind and suddenly throw
> the COST-based optimizer at it. Whilst perhaps tuned perfection on the
> other, all sorts of weird and wonderful execution plans result when the
> optimizer mode suddenly changes like that.
What about newer schemas that weren't around before the CBO - like
OUTLN and DBSNMP. Is it safe to gather statistics on them?
Received on Fri Feb 28 2003 - 11:17:58 CST
