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: Oracle Myths

Re: Oracle Myths

From: Pablo Sanchez <pablo_at_dev.null>
Date: Wed, 15 May 2002 14:57:46 -0600
Message-ID: <3ce2ca11$2_16@news.teranews.com>

"Daniel Morgan" <dmorgan_at_exesolutions.com> wrote in message news:3CE2A78A.79A3290B_at_exesolutions.com...
> Alan wrote:
>
> > Can someone explain why "Seperate tables and indexes for
performance
> > reasons." is not true?
>
> To avoid your being pummeled ... let me suggest that you look in the
> Google archives. This one was painfully beaten into submission a few
> weeks ago. And as I score it ... the experts convinced the skeptics.

Wow, I missed that thread ... anyway, I separate the two for one and only one reason:

    Different I/O characteristics between the two types of data.

I find that typically there's some locality of data in 'data' versus 'indexes' which is why I love DB_FILE_MULTIBLOCK_READ. Sure we might do an index scan but typically not.

--
Pablo Sanchez, High-Performance Database Engineering
mailto:pablo_at_hpdbe.com
http://www.hpdbe.com
Available for short-term and long-term contracts
Received on Wed May 15 2002 - 15:57:46 CDT

Original text of this message

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