X-Original-To: pgsql-advocacy-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id 4CD8E533F7 for ; Mon, 28 Mar 2005 18:05:53 +0100 (BST) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 11640-02 for ; Mon, 28 Mar 2005 17:05:49 +0000 (GMT) Received: from wolff.to (wolff.to [66.93.249.74]) by svr1.postgresql.org (Postfix) with SMTP id 0EC685337A for ; Mon, 28 Mar 2005 18:05:48 +0100 (BST) Received: (qmail 11656 invoked by uid 500); 28 Mar 2005 17:22:23 -0000 Date: Mon, 28 Mar 2005 11:22:23 -0600 From: Bruno Wolff III To: Virgile Beddok Cc: Tom Lane , Christopher Kings-Lynne , pgsql-novice@postgresql.org, pgsql-advocacy@postgresql.org Subject: Re: [NOVICE] Function which gives back the nearest neighbours Message-ID: <20050328172223.GA11567@wolff.to> Mail-Followup-To: Bruno Wolff III , Virgile Beddok , Tom Lane , Christopher Kings-Lynne , pgsql-novice@postgresql.org, pgsql-advocacy@postgresql.org References: <2657.146.140.8.144.1111775776.squirrel@webmail.igd.fraunhofer.de> <42464392.9040903@familyhealth.com.au> <20050327163653.GH9280@wolff.to> <5796.1111941132@sss.pgh.pa.us> <1218.146.140.8.94.1112025141.squirrel@webmail.igd.fraunhofer.de> <20050328161240.GD7217@wolff.to> <1343.146.140.8.94.1112027529.squirrel@webmail.igd.fraunhofer.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1343.146.140.8.94.1112027529.squirrel@webmail.igd.fraunhofer.de> User-Agent: Mutt/1.5.6i X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.011 tagged_above=0 required=5 tests=AWL X-Spam-Level: X-Archive-Number: 200503/123 X-Sequence-Number: 6551 On Mon, Mar 28, 2005 at 18:32:09 +0200, Virgile Beddok wrote: > > I just want to know if there is an existing and implemented function, or > tree, in Postgres which allows me to directly perform a "nearest neighbour > search" on multidimensional vectors. How are you measuring distance? (i.e. Euclidean distance isn't the only metric that you could be using.) Earthdistance has functions for doing this using geodesics on the surface of a sphere. For 2D, I think there is a function on "point"s that uses Euclidean distance. I don't think this is the hard part of your problem. It is easy to create a metric function and get the point corresponding to the minimal value using a sort. Being able to limit the sets of points looked at is the hard part.