Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Index management

Re: Index management

From: Mike Ault <mikerault_at_earthlink.net>
Date: 11 May 2004 05:28:35 -0700
Message-ID: <37fab3ab.0405110428.38ef171e@posting.google.com>


Howard,

Your graciousness only eclipses your tactfulness. At sites with tables that contain millions (in one or two cases, billions) of rows doing a compute rather than estimate on statistics is not realistic or possible. Had you the experience that your 30 years in the field should have granted you this would be obvious.

You also assume that the Oracle system is bug free and always does everything correctly. I can't afford to assume that as my experience has shown otherwise. Shoot, they can't even get simple row counts correct much of the time.

I understand what the absolute clustering factor is, and also understand that there are no absolutes in the real world, everything must be tested as algorithms are implemented incorrectly on a daily basis across much of the industry.

As far as I am concerned, the tread is closed.

Mike Received on Tue May 11 2004 - 07:28:35 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US