Re: 7.23 UC4 vs. 7.30 UC3

In article <35FB8803.4217@cyberramp.net>, greg@cyberramp.net wrote: > snip... > > >> Also, Oyvind, we have experienced the wrong query path being selected by the > > >> optimizer is 7.x for some time. We understand that 7.30 allows the user (us) > > >> to control the query optimizer more directly. > > > > > >Yes, we have used this feature to verify that the optimizer chose the > > >wrong query path. The cost was _much_, _much_ lower when we suggested the > > >correct index. However this is something I would hesitate to use in a > > >production environment. It is extremly painful to maintain. > snip.... > > I've done a lot of performance tuning, and most of the time when the > optimizer chooses the wrong path its because it needs better statistics > to push it in the right direction. Have you explored this angle yet? > Its really preferable to hardwiring hints into your code. > > Greg > Good point Greg. To do this I'll have to sit down determine what is to be done when. Some of our tables are large and rather dynamic. How do you handle scheduling of update statistics on your tables? I already know about statistics versus distribution, and that this can be done at the column level. Did you have to do a lot of analysis and planning, and how did you manage this? -----== Posted via Deja News, The Leader in Internet Discussion ==----- http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum