Re: SW vendor insists on RBO

Niall,

You don't have to convince me of the desirablity to use CBO. I offered our vendor to do part of the analysis if problems arise when changing from RBO to CBO.
Of course it's bad design, and they know it, and theuy agree it should be fixed, but as things go, our vendor doesn't want to invest much time (at this moment, so they say) in certifying their app against CBO.

My post was meant to find out if one might expect more unexpected errors (because of bad design of course ;-) ), comparable to the ORA-1722 we had. I *was* expecting differences with respect to performance, but received weird errors in stead.

Peter

"Niall Litchfield" <niall.litchfield_at_dial.pipex.com> schreef in bericht news:3db1c2e1$0$13370$cc9e4d1f_at_news.dial.pipex.com...
> "Peter van Rijn" <pgm.van.rijnRM_at_THIShccnet.nl> wrote in message
> news:aoq2l7$a7l$1_at_news.hccnet.nl...
> > Now we're getiing to my point. I really intended to use the CBO, even
when
> > our SW vendor advised against it, I'm pretty sure 8.1.7 CBO is usually
> > better than RBO. But now I have my doubts. What other surprises can I
> > expect? I never would have thought I'd encounter software problems
because
> > of using another optimizer.
>
> In this case it seems to me that you are hitting software problems not
> because of the optimiser in use, but because of a bad design decision (to
> store inconsistent datatypes in the same column). In my book your app
should
> either be treating the strings as strings OR storing numbers in a number
> field.
>
> Other things you should consider are the long list of new features that
are
> not supported by the RBO and so either will not be used effectively (or
will
> kick of the CBO), and wether the advice fronm the vendor is 'advice' or a
> support requirement.
> --
> Niall Litchfield
> Oracle DBA
> Audit Commission UK
> *****************************************
> Please include version and platform
> and SQL where applicable
> It makes life easier and increases the
> likelihood of a good answer
> ******************************************
>
>
>
Received on Sun Oct 20 2002 - 04:20:53 CDT