Re: Statistics based optimizer usable?

From: Clay Jackson <cjackso_at_uswnvg.com>
Date: 9 May 94 20:21:35 GMT
Message-ID: <31466_at_uswnvg.uswnvg.com>


Georg Rehfeld (georg_at_deerwood.hanse.de) wrote:
: . .-. +---- +----
: /| | | | |
: | | | +--. +--.
: | | | ) )
: --- `-' ---' ---' seconds (more than 15 minutes to PARSE !!!!!!!)
 

: while the 'execute' and 'fetch' times are in good range (0.01 and 0.03 sec).
 

: - has one of you experienced similar results?
Interesting - yes, we've seen the same sort of problems with 7.0.15.4 on a Sequent box. We didn't really spend any time on it, because it was an error on our part to not have 'OPTIMIZER=RULE' in our init.ora (we've decided to ONLY try the cost-based optimizer on a limited basis.

: - has on of you a solution or suggestion?
Use the rule based optimizer, at least it's predictable.

: - has one of you used the statistic based optimizer with good results?
Not yet - I'll post something if we do.

: - are there known bugs/limitations in using the statistic based optimizer?
I hope not...

--
Clay Jackson - N7QNM
US WEST NewVector Group Inc
Bellevue, WA
uunet!uswnvg!cjackso
Received on Mon May 09 1994 - 22:21:35 CEST

Original text of this message