Re: full-text indexing

On Tue, 18 Apr 2000, Bruce Momjian wrote:

> I have one word for you:  CLUSTER.  Without it, index lookups are too
> slow.  With it, they are rapid.  I have done some work like this
> commerically with Ingres, which has an ISAM type that keeps the matching
> rows pretty close on a newly-created ISAM index.  In PostgreSQL, and
> regular CLUSTER will keep you good.

I agree!  The last bit of advice given in the full text README.  As I
said, I'd built full-text stuff for experimentation (I had maybe 30k of
raw text, which amounted to several 100,000 indexed entries), and I had
clustered it, and it was pretty darn fast, even on a Pentium 233 with only
48 megs of RAM.  I have significantly better hardware to run it on now.
The original project called MySQL, but it just didn't have what we needed
to put something like this together.

> If you find it slow, let me know.  I have done some benchmarking with
> the author and he found it pretty fast, usually a few seconds.  See the
> section in my book on CLUSTER for information on _why_ it helps.

Thanks, Bruce.

Brett W. McCoy
                                              http://www.chapelperilous.net
---------------------------------------------------------------------------
Twenty two thousand days.
Twenty two thousand days.
It's not a lot.
It's all you've got.
Twenty two thousand days.
        -- Moody Blues, "Twenty Two Thousand Days"