Re: Oracle db optimizer
Date: Tue, 20 May 2014 07:06:46 -0700 (PDT)
Message-ID: <da400539-f3af-412a-b852-c66346d195e9_at_googlegroups.com>
On Tuesday, May 20, 2014 3:40:10 AM UTC-6, Jack wrote:
> "joel garry" wrote in message
>
> news:609a4cc1-fd8c-4a1d-88aa-113c7d5df5f5_at_googlegroups.com...
>
>
>
> On Monday, May 19, 2014 5:59:52 AM UTC-7, Jack wrote:
>
> > "Jack" wrote in message news:sshev.40752$SH2.19490_at_uutiset.elisa.fi...
>
> >
>
> >This isn't really enough information to say, but for whatever reason your
>
> >BS table doing a full scan is a much better plan than a bunch of >nested
>
> >loops. One could guess the statistics or lack of statistics in the XE is
>
> >responsible for that, but better to know than guess. Here is one >simple
>
> >methodology for giving us enough information to tell - and you may work it
>
> >out just following the steps. Ignore the formatting >advice, of course.
>
> >http://oracle-randolf.blogspot.com/2009/02/basic-sql-statement-performance.html
>
> >
>
> >This all assumes you've gotten the basics from the performance tuning
>
> >manual. Jonathan Lewis has a book that explains it in detail, highly
>
> > >recommended.
>
> >
>
> >jg
>
> >--
>
> >_at_home.com is bogus.
>
> >http://www.microscope.co.uk/news/2240220895/Oracle-extends-reach-with-desktop-virtualisation-specialist
>
>
>
> Thanks jg for your comment.
>
>
>
> Not sooooo interested to read 27 pages report about plan.
>
> Well, no hint available to fix it. (or _dbparameter ?)
>
> So I made the logic more simple. Leaving some details away.
>
> Some exceptions (and extraordinary exceptionals) are now not counted.
>
>
>
> But hmm. db 11's costbased optimizer looks very good. No glitches found.
>
> Yet.