Re: Building indexes in parallel

From: Jim Olsen <jolsen_at_atl.mindspring.com>
Date: 1996/04/23
Message-ID: <4lil0l$i0k_at_mule1.mindspring.com>#1/1


Try
create index index_name on table_name(column_name) tablespace ts_name
parallel (degree 4)
storage(initial size next size);

Be careful that if you say initial 10M and degree of 4 it will create for extents for this segment and make each 10M for a total of 40M for your index.

You may also wish to include the unrecoverable option anywhere in the create index statement. This option will create the index without writing any redo-log inormation. This will speed up your index build.

duelund_at_inet.uni-c.dk (Niels Duelund) wrote:

>Hi there
 

>I am using an AIX, multiprocessor system (RISC6000/R30) with 4 CPU's.
 

>Does anybody know about building indexes in parallel. Meaning
>using 3-4 streams for building one index ??

>Thanks in advance
> Niels Duelund (Duelund Data, Denmark)
Received on Tue Apr 23 1996 - 00:00:00 CEST

Original text of this message