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: Creating indexes on a large table

Re: Creating indexes on a large table

From: Mladen Gogala <gogala_at_sbcglobal.net>
Date: Fri, 05 Aug 2005 11:35:13 GMT
Message-Id: <pan.2005.08.05.11.35.12.790621@sbcglobal.net>


On Mon, 01 Aug 2005 13:54:39 -0700, prmail_at_gmail.com wrote:

>
> Creating indexes on a large table
>
> Hi,
> I have 2 large tables on which we have to create indexes.
>
> Table 1: ( Contains 4 Billion Records )
>
> Col 1 : Col 2 ( Datatype Number (10) )
>
> Table 2 ( Contains 1 Billion Records )
>
> Col 3 : Col 4 : Col 5 ( Datatype Number (10) )
>
> What is the most efficient way to create indexes on all the columns. I
> would be using all the columns separately in queries.

NOLOGGING PARALLEL <as much as possible>;

Make sure that your temporary tablespace is large. No, even larger then that: humongous. Don use auto management of workarea, set SORT_AREA_SIZE to 128M (be aware that each of P000n processes will use that amount of memory) and hope for the best. If your database is RAC, use all CPU resources, across the RAC by specifying an appropriate degree of parallelism. May the force be with you.

-- 
http://www.mgogala.com
Received on Fri Aug 05 2005 - 06:35:13 CDT

Original text of this message

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