Re: Update Statistics Plan for IDS 7.31

Gack! HTML post! Yech! From: Greg Dewinter <gdewinter@spf.fairchildsemi.com> ><!doctype html public "-//w3c//dtd html 4.0 transitional//en"> ><html> >&nbsp; ><br>Obnoxio The Clown wrote: ><blockquote TYPE=CITE>From: "Art S. Kagel" &lt;kagel@bloomberg.net> ><br>> ><br>>Asside to Obnoxio:&nbsp; You guys were doing OK.&nbsp; What'd you >need me for?&nbsp; ;-) ><p>If you have two indexes on a table, one on (a,b,c) and one on >(a,b,d),</blockquote> >To start I would ask why you wouldn't create the indexes on in the >following ><br>manner instead: ><br>(a,b,c) ><br>(d,b,a) - or switch b and d This is a bit of an artficial and simplified example. You may have a specific usage pattern that "recommends" the indexes be in similar order... ><br>other that work one one database and not on another.&nbsp; Too >summarize, ><br>each database is different, so why would the same statistics scheme >work on ><br>every database the same way?&nbsp; I guess I really didn't answer Because the same optimiser is working on all the databases? :-) ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com