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 -> Building large indices

Building large indices

From: Greg Stark <greg-spare-1_at_mit.edu>
Date: Wed, 20 Oct 1999 15:32:53 GMT
Message-ID: <87r9iqt5cs.fsf_-_@mit.edu>

"Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> writes:

> Oracle 8.1.5 allows on-line rebuild of indicesm otherwise
> no. For earlier versions an 'unrecoverable' create index
> in parallel is probably the fastest thing to do.

There's no way to take advantage of the existing index which is a prefix of the new one to build the new one faster?

For another table where I want to index entirely new columns would it be faster to create the index with the columns all nulls? I don't care how long it takes to actually fill out the columns as long as the downtime for creating the index is minimized.

On this note, about how long should it be taking to build an index on a million records anyways? On a reasonably optimized dual processor sparc should it be 1 minutes or 1 hour? I realize this depends a lot on the hardware but I'm just trying to get a feeling for how bad the outage will be and just whether something is seriously wrong with the way the database is running currently.

--
greg Received on Wed Oct 20 1999 - 10:32:53 CDT

Original text of this message

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