Another solution is to use a knngist index. It will most likely be incorporated in a future version of PostGIS. Since it looks like it was already commited in PostgreSQL trunk, and if you have one of those brave-trunk-running souls, you should be able to test it with a beta version of PostgreSQL. You will get a considerable speed gain from using this approach.
- Ragi Date: Thu, 24 Feb 2011 20:38:36 -0800 (PST) > From: Scholle <mstu...@gmail.com> > Subject: Re: [postgis-users] Find n Nearest Neighbors for given Point > using PostGIS? > To: postgis-users@postgis.refractions.net > Message-ID: <31010203.p...@talk.nabble.com> > Content-Type: text/plain; charset=us-ascii > > > Great, didn't consider the geometry/degree difference.... I drastically > decreased the value for the third parameter of ST_DWithin function and its > sufficiently fast now... > > > > > Ben Madin-3 wrote: > > > > Have you tried EXPLAIN to see where the slow part is? > > > > But at a guess - consider that st_dwithin uses the geometry unit for it's > > calculations - so you are searching for everything within 300 degrees > > (more than halfway around the planet). You may want to try searching a > > smaller set of data before you sort it to find the closest five. > > > > cheers > > > > Ben > > > > On 25/02/2011, at 12:04 PM, Scholle wrote: > > > >> > >> I am trying to solve the problem of finding the n nearest neighbors > using > >> PostGIS: > >> > >> Starting Point: > >> > >> - Table geoname with geonames (from geonames.org) containing > >> latitude/longitude (WSG-84) > >> - Added a GeometryColumn geom with srid=4326 and datatype=POINT > >> - Filled geom with values: UPDATE geoname SET geom = > >> ST_SetSRID(ST_Point(longitude,latitude) 4326); > >> - Created GIST index for geom (CREATE INDEX geom_index ON geoname USING > >> GIST (geom);) / Clustered geom_index: CLUSTER geom_index ON geoname;) > >> - Created PRIMARY KEY UNIQUE BTREE index for geonameid > >> > >> Problem: > >> Find n (e.g. 5) nearest neighbors for a given Point in table geoname > >> represented by id (geoname.geonameid. > >> > >> Possible solution: > >> > >> Inspired by > >> > http://www.bostongis.com/PrinterFriendly.aspx?content_name=postgis_nearest_neighbor > , > >> I tried the following query: > >> > >> "SELECT start.asciiname, ende.asciiname, distance_sphere(start.geom, > >> ende.geom) as distance " + > >> "FROM geoname As start, geoname As ende WHERE start.geonameid = 2950159 > >> AND > >> start.geonameid <> ende.geonameid " + > >> "AND ST_DWithin(start.geom, ende.geom, 300) order by distance limit 5" > >> > >> Processing time: about 60s > >> > >> Also tried an approach based on EXPAND: > >> > >> "SELECT start.asciiname, ende.asciiname, distance_sphere(start.geom, > >> ende.geom) as distance " + > >> "FROM geoname As start, geoname As ende WHERE start.geonameid = 2950159 > >> AND > >> start.geonameid <> ende.geonameid AND expand(start.geom, 300) && > >> ende.geom " > >> + > >> "order by distance limit 5" > >> > >> Processing time: about 120s > >> > >> The intended application is some kind of autocomplete. So, any approach > >> taking longer than <1s is not applicable. Is it generally possible to > >> achieve such a response time with PostGIS? > >> -- >
_______________________________________________ postgis-users mailing list postgis-users@postgis.refractions.net http://postgis.refractions.net/mailman/listinfo/postgis-users
