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: How to tune a huge SELECT ?

Re: How to tune a huge SELECT ?

From: Nuno Souto <nsouto_at_optushome.com.au>
Date: 12 Sep 2002 20:54:20 -0700
Message-ID: <dd5cc559.0209121954.f315e9b@posting.google.com>

chse30_at_hotmail.com (Christian Svensson) wrote in message news:<ccc2a7eb.0209060550.55afae93_at_posting.google.com>...
>
> Since these SELECT get a lot of data, we are talking about approx 100
> million rows. It takes approx 7-8 hours for Cognos to select this data
> into a Cognos tempfile where it later use this to build the cubes.

100 million rows is a LOT of rows in anyone's language. Nor surprised it takes that long to create the file! I'm not even going to ask how long it takes to build the cubes!

>
> What I wonder is what Oracle setting/tuning can you make when you want
> to get this amount of data from the database ?

Hmmmm,
_NETWORK_FAST = TRUE
comes to mind. Another one I'd try would be: _PARALLEL_EXECUTION = TRUE. _LOTS_OF_CPUS = TRUE
rings a bell as well.

;)

Might look like I'm pulling your leg, but read them again. They mean what you got to look at, not some magical silver bullet that will in a single stroke drop those times to minutes.

> I am open for any suggestions.
>

remain open. That's the key for innovative approaches to this problem.

Cheers
Nuno Souto
nsouto_at_optushome.com.au Received on Thu Sep 12 2002 - 22:54:20 CDT

Original text of this message

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