Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: Rule Based Optimization Going Away?
Could someone post the actual article from metalink? I don't have a
metalink account. I'd like to read it with my own eyes.
thanks,
-Peter Kellner
"Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message news:<7ZMm9.45097$g9.128606_at_newsfeeds.bigpond.com>...
> "Peter Kellner" <pkellner_at_seventythirdstreet.com> wrote in message
> news:ccec98ba.0210021655.32903b35_at_posting.google.com...
> > I just was told by an Oracle Consultant that Oracle is planning on
> > obsoleting rule based optimization and that in the future sometime,
> > everyone will have to use cost based. Could this be true?
>
> Yup. There's a public document admitting as much on Metalink.
>
> >I have
> > always used rule based when I know the distribution of the data pretty
> > well ahead of time and can be smarter than the cost optimizer.
> >
>
> Then you need to get with the rhythm and realise that the Opimizer has got a
> whole lot more smarter than you're giving it credit for. So when was the
> last time you could predict, with the rule based optimizer, what would be in
> Index leaf nodes before you actually got to read them??? Huh??
>
> Seriously: there is no way you can beat the optimizer these days. And if you
> say you can, I don't believe you.
>
> Anyway, whether I believe you or not: by 10i, the rule-based optimizer is
> finished. So even if you are the God's gift to predicting access paths,
> tough luck: Oracle won't support you.
>
> > Any official Oracle Info on this would be appreciated.
> >
>
> You don't need any more "official" info than is readily available from
> Metalink. The Rule Based Optimizer is DEAD and you need to upgrade. It's not
> a particular pain: the CBO is good these days. But you need to stop wearing
> the flared jeans and oversized-lapels, and get with the groove. Man.
>
> Regards
> HJR
>
>
> > -Peter Kellner
Received on Thu Oct 03 2002 - 00:25:51 CDT
![]() |
![]() |