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: Myth revisited ...

Re: Myth revisited ...

From: Niall Litchfield <niall.litchfield_at_dial.pipex.com>
Date: Sun, 16 Nov 2003 20:05:10 -0000
Message-ID: <3fb7d87d$0$3347$cc9e4d1f@news.dial.pipex.com>


"Noons" <wizofoz2k_at_yahoo.com.au.nospam> wrote in message news:3fb74440$1$13634$afc38c87_at_news.optusnet.com.au...
> There was a thread on Oracle-L about this a few weeks
> ago. Funny how these things come around.
>
> The consensus seemed to be you can get a measure
> of better indexing (less b-tree levels) if you put
> the index in a tablespace with a large block size.
> Therefore reducing I/O even more for the indexes.

http://www.ixora.com.au/tips/block_size.htm

has a pretty good discussion on large block size and indexed access paths. I'm not sure how realistic it would be to expect b-level reduction for most indexes but for range scans and the like the argument would seem to hold well.

> As for tables in tablespaces of larger block sizes,
> I can see an advantage when dealing with LOBs: you get
> (hopefully) more "in-lined" rows. It also may be
> advantageous for IOTs.

I'm not sure that I necessarily think that in-line LOBS are *ever* a good idea. ISTM that in nearly all the cases where an inline LOB would be appropriate, then a Varchar field would probably be a better thing to do.

--
Niall Litchfield
Oracle DBA
Audit Commission UK
*****************************************
Please include version and platform
and SQL where applicable
It makes life easier and increases the
likelihood of a good answer
******************************************
Received on Sun Nov 16 2003 - 14:05:10 CST

Original text of this message

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