Re: performace question...

Vince Laurent <eAddict_at_yahoo.com> wrote:
> Quick question: Are full table scans bad for performance? I am
> guessing yes. I ran this script:

When it should be doing something other than FTS, then doing FTS is bad for performance.

>
> set serverout on size 1000000
> set verify off
>
> col object_name form a30
>
> PROMPT Column flag in x$bh table is set to value 0x80000, when
> PROMPT block was read by a sequential scan.
>
> spool bufferts.lst
>
> SELECT o.object_name,o.object_type,o.owner
> FROM dba_objects o,x$bh x
> WHERE x.obj=o.object_id
> AND o.object_type='TABLE'
> AND standard.bitand(x.flag,524288)>0
> AND o.owner<>'SYS';
>
> spool off
>
> and the result on one of our DB was 5600!

5600! is about 10^18560. That is one mighty big cache you have there.

So what is it that you think you've just calculated with this query? That you've performed 5600 FTS? That you performed one FTS that touched 5600 rows? That 5600 blocks have been part of a FTS?

And whatever you think it means, why do you think that whatever you think it means is bad for performance?

> Since Sunday's backup! How
> can I tell the vendor that thier appliction, um, er... works harder
> than it needs to?

Well, the telephone is a good way to go. But it always helps if you know what you are talking about first.

Xho

-- 
-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service                        $9.95/Month 30GB