Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> c.d.o.server -> Re: Reduce or shrink the physical size of a text index

Re: Reduce or shrink the physical size of a text index

From: Snewber <>
Date: Wed, 07 Sep 2005 15:21:24 +1000
Message-ID: <dfltck$cbh$>

Thanks for the information. Unfortunately I was consulted about the text indexing and so I now have to review all the parameters to fix things up. I have to review case insensitive options, no indexing of numbers, the stemming stuff that includes words of 3 to 6 characters and also the addition of extra stopwords.

Yes it is the syncrhonising that is causing the blow outs. I just dropped and recreated the indexes and this resulted in their sized dropping from gigabytes to a few hundred Mb. I have also implemented index optimisation first to reduce index size.

What do you mean by compress? Guess I'll look into this.

Also, the doco states not to use the alter index ... stuff for context indexes, but the alter index ... rebuild ... parameter ... seems to be the only way to reduce the size of the indexes online, although I couldn't get the parameter syntax to work correctly so I'm not sure if this would reduce physical size?

Pratap wrote:
> 1. May be you sync your index too often. In that case use should
> optimize the index from time to time (See API for optimization). You
> can do this online and this will reduce the size of the index by
> eliminating the duplicate tokens in the $I table
> 2. Use compress 2 for $I indexes. Don't fire the alter on the $I table
> directly. See the docs for how to do this for text indexes
> 3. See if your stoplist is intact. You may wish to introduce more noice
> words in the stoplist by analyzing your data (You won't be able to
> search using these words though)
> Regards,
> Pratap
Received on Wed Sep 07 2005 - 00:21:24 CDT

Original text of this message