Re: Index management

"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:409d7a18$0$442$afc38c87_at_news.optusnet.com.au...

>>Mike Ault wrote:

> Well, it appears we are never going to get an explanation of what a
> dirty base table block is (I've only asked three times, and Daniel's
> asked too).
>
> But whatever this mysterious beastie happens to be, what possible
> difference can it make to such a ratio to rebuild an index??
>

Please, please Mike, add me to the list of requestors for this explanation. Hopefully a clear definition of "dirty base table block" will add more light than heat to this discussion.
>
> > By definition the order of rows in a relational table is random.
>
> Also absolutely true.
>

Howard, are you sure of this? Relational theorists please correct me if I am wrong, but I thought the best we could say about the order of rows is that it is indeterminate, which means we can't even make statistical arguments based on assumed randomness.
>
> HJR
My only experience with this was with a third party vendor who slipped a weekly "rebuild every index in the system" job into a production system I was babysitting without saying a word to anyone (Grrrr!) So I can add a note here that index rebuilds of any real size are expensive in both CPU and log archive space as well as being almost always ineffective and pointless.

Thank you folks for an informative and so far entertaining thread. I, for one, promise faithfully never ever to rebuild an index (not that I ever would or did, mark you) without revisiting this thread and rereading every post in it.

Roger S Gay Received on Sat May 08 2004 - 20:49:52 CDT