tag:blogger.com,1999:blog-5125852724824145995.post6672299323274202482..comments2016-08-12T06:46:32.724-07:00Comments on Geogeek: Voronoi Diagrams in PostGISGeoGeekhttp://www.blogger.com/profile/02596534612535469564noreply@blogger.comBlogger8125tag:blogger.com,1999:blog-5125852724824145995.post-50147812298130292832012-09-11T08:43:13.343-07:002012-09-11T08:43:13.343-07:00After trying the plpythonu Voronoi function and pr...After trying the plpythonu Voronoi function and producing artefacts and overlapping and self-overlapping voronoi polygons all the time (same happens if you use the ftools voronoi tool in QGIS, as the plpython code is derived from this), I tried out the PL/R voronoi function from http://punkish.org/Voronoi-Diagrams-In-PostGIS and found out that after some small adaptions to postgis 2 (make srid()-->st_srid() etc), the function produced a flawless set of voronoi polygons for 1000 points in about 13 seconds on a Core2duo CPU. raliskihttp://www.blogger.com/profile/11828343425764159849noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-38209130414807345812012-08-22T08:54:58.790-07:002012-08-22T08:54:58.790-07:00Hi Paul,
The code is available. See http://geogee...Hi Paul,<br /><br />The code is available. See http://geogeek.garnix.org/2012/04/faster-voronoi-diagrams-in-postgis.html<br /><br />If you look through the comments you'll see that it only works with Python 2.x – the 3.x version still has some bugs I haven't had time to fix.<br />GeoGeekhttp://www.blogger.com/profile/02596534612535469564noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-24552109486568605732012-08-20T15:19:18.071-07:002012-08-20T15:19:18.071-07:00Was wondering if you could make your modified code...Was wondering if you could make your modified code available? I've come across the other Voronoi example on BostonGIS but it dependso n PL/R and the R stats packages. I'd like to work with your example as it sounds like it doesn't have that dependency.<br /><br />ThankPaul Wickmanhttp://www.blogger.com/profile/17311942878098442838noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-82827955234514769752012-04-05T12:03:53.193-07:002012-04-05T12:03:53.193-07:00OK, I did exactly that. I'll have a post up on...OK, I did exactly that. I'll have a post up on it shortly. I just did a test on 7900 points and it took 7.5 seconds, and 40,000 points took 20 seconds.GeoGeekhttp://www.blogger.com/profile/02596534612535469564noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-22081088333933681782012-04-05T08:46:18.782-07:002012-04-05T08:46:18.782-07:00It's true it's not the most space efficien...It's true it's not the most space efficient algorithm. The only suggestion I can think of off the top of my head is to do it "batches" of a thousand or so, grouped by the_geom, and then merge the output of each batch together somehow.<br /><br />The algorithm is pretty naïve. I suppose it wouldn't be too challenging to implement Fortune's algorithm, but I'd probably want to modify an existing implementation (such as this one from QGIS: https://svn.osgeo.org/qgis/trunk/qgis/python/plugins/fTools/tools/voronoi.py) to work inside postgres.<br /><br />That actually doesn't look like too much of a challenge, and if I get a chance, I might try to do that.GeoGeekhttp://www.blogger.com/profile/02596534612535469564noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-40041072718507086532012-04-02T04:17:10.601-07:002012-04-02T04:17:10.601-07:00Hey I am trying this function for a dataset with 1...Hey I am trying this function for a dataset with 10.000 points, it fills up more than 50 Gigabytes of disk space and then is cancelled because lackof disk space. Is there any idea how to decrease the disk psace use for this function a bit?<br />Thanks!rayhttp://www.blogger.com/profile/13777772035471547485noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-16387930016566594452011-02-21T07:25:04.248-08:002011-02-21T07:25:04.248-08:00Hi!
This is awesome! I'm very glad to see a f...Hi!<br /><br />This is awesome! I'm very glad to see a free option to the ESRI Voronoi utility. Works very well, except I've noticed that sometimes, polygons are missing from the results. Any idea why that happens?<br /><br />Thanks!GraniteLibertyhttp://www.blogger.com/profile/15935048064283735090noreply@blogger.comtag:blogger.com,1999:blog-5125852724824145995.post-39156315333735524282011-01-27T01:48:01.500-08:002011-01-27T01:48:01.500-08:00hello!!
thanks for your script! I have been using ...hello!!<br />thanks for your script! I have been using it successfully on postgres 9.0. I just had to modify the script using the st_union() function instead of the geomunion()one. Hope it will be usefull for you!<br />Sorry for my english, I'm a french customer.anaishttp://www.blogger.com/profile/04789810348614738423noreply@blogger.com