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: Noons <wizofoz2k_at_yahoo.com.au>
Date: Sun, 2 May 2004 21:59:12 +1000
Message-ID: <4094e4c7$0$12033$afc38c87@news.optusnet.com.au>


Interesting. In which version have you found this? A long time ago Oracle had an asynch process that performed read aheads for table scans. In V4, IIRC. It went away with V6 and DMBR replaced it. I think at some stage there was talk of it being re-introduced, around the 9i release, for any long scan. Not sure if it ever was or if it fell away. This might be a left-over from that idea. Makes sense too.

-- 
Cheers
Nuno Souto
wizofoz2k_at_yahoo.com.au.nospam
"Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> wrote in message
news:c70ngi$r20$1_at_sparta.btinternet.com...
> 2)  Oracle has an 'index prefetch' algorithm

> that I believe is supposed to kick in for large
> index range scans. I think I've seen it happen
> a couple of times on an index FULL scan (NB
> I don't mean 'fast full scan'). I think the code
> works on the assumption that the leaf blocks
> for a large range scan are likely to be adjacent,
> so a read request is made for multiple blocks,
> and irrelevant ones are discarded.
Received on Sun May 02 2004 - 06:59:12 CDT

Original text of this message

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