To: tjo@acm.org
Cc: pgsql-sql@postgresql.org
Subject: Re: several questions about R-tree index 
In-reply-to: <3366.209.223.166.104.1114559525.squirrel@gnova.com> 
References: <3366.209.223.166.104.1114559525.squirrel@gnova.com>
Comments: In-reply-to "TJ O'Donnell" <tjo@acm.org>
	message dated "Tue, 26 Apr 2005 16:52:05 -0700"
Date: Wed, 27 Apr 2005 00:12:32 -0400
Message-ID: <17645.1114575152@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

"TJ O'Donnell" <tjo@acm.org> writes:
> Shouldn't the ~ (contains) operator be included also?
> Isn't ~ the commutator of @ ?

Yeah, it looks like the documentation is in error:

regression=# select amopopr::regoperator,amopstrategy from pg_amop where amopclaid in
regression-# (select oid from pg_opclass where opcamid = 402);
       amopopr       | amopstrategy 
---------------------+--------------
 <<(box,box)         |            1
 &<(box,box)         |            2
 &&(box,box)         |            3
 &>(box,box)         |            4
 >>(box,box)         |            5
 ~=(box,box)         |            6
 ~(box,box)          |            7
 @(box,box)          |            8
 <<(polygon,polygon) |            1
 &<(polygon,polygon) |            2
 &&(polygon,polygon) |            3
 &>(polygon,polygon) |            4
 >>(polygon,polygon) |            5
 ~=(polygon,polygon) |            6
 ~(polygon,polygon)  |            7
 @(polygon,polygon)  |            8
(16 rows)

Of course, the thing that leaps out here is that there are only two
built-in opclasses for rtree.  Not exactly confidence building.

> I am considering using R-tree's for other-than-geometric purposes.

I concur with Andrew's suggestion to consider GIST.  GIST has its own
issues, but at least there are people looking at it/using it/working on it.
R-tree doesn't seem to have any user community that really cares.

			regards, tom lane