Re: [postgis-users] curves geometries and topology
Thanks! We are currently considering using PostGIS Topology only to maintain a topology aware copy of the data (to be used for analyses where topology matters) and use the original geometry (containing circular strings) when we have to (re)generate output geometry. BTW: Is cgal/sfcgal going to replace geos eventually? It would be great if this limitation (= curved geometries being a second class citizen) was somehow lifted in future versions PostGIS. Regards, Reijer Copier IDgis bv Volg IDgis op LinkedIn! http://www.linkedin.com/company/idgis On 14-4-2014 19:15, Rémi Cura wrote: The problem is going to be precision. If this is important for you you may want to switch to cgal which support arbitrary curve in arrangment (from memory) If precision is not important, you can easily add an uggly fix : add 2 triggers to edge table that _when something want to access an edge, convert it to linestring _when something want to write an edge, convert it to your curve model Cheers, Rémi-C 2014-04-14 17:10 GMT+02:00 Eric Ladner eric.lad...@gmail.com mailto:eric.lad...@gmail.com: As long as they are simple circles, you should be able to convert back and forth between curves and lines. Splines, beziers, et al, are probably not going to work at all and would need to be approximated as lines. http://boundlessgeo.com/2012/01/getting-curvey/ On Mon, Apr 14, 2014 at 8:02 AM, Reijer Copier reijer.cop...@idgis.nl mailto:reijer.cop...@idgis.nl wrote: Dear list, According to the documentation, PostGIS Topology only supports edges made out of linestrings. However, we are currently working on a dataset that contains lots of curved geometries. It would be great if we could build topology based on those curved geometries without having to convert them to linestrings first. Is this possible? If not, what is the best way to deal with this situation? -- Kind regards, Reijer Copier IDgis bv ___ postgis-users mailing list postgis-users@lists.osgeo.org mailto:postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users -- Eric Ladner ___ postgis-users mailing list postgis-users@lists.osgeo.org mailto:postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] curves geometries and topology
On Tue, Apr 15, 2014 at 04:40:34PM +0200, Reijer Copier wrote: Thanks! We are currently considering using PostGIS Topology only to maintain a topology aware copy of the data (to be used for analyses where topology matters) and use the original geometry (containing circular strings) when we have to (re)generate output geometry. BTW: Is cgal/sfcgal going to replace geos eventually? It would be great if this limitation (= curved geometries being a second class citizen) was somehow lifted in future versions PostGIS. It could be, once it gets to support everything currently supported by the GEOS backend and proves to be better maintained :) --strk; ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] curves geometries and topology
2014-04-15 16:58 GMT+02:00 Sandro Santilli s...@keybit.net: On Tue, Apr 15, 2014 at 04:40:34PM +0200, Reijer Copier wrote: Thanks! We are currently considering using PostGIS Topology only to maintain a topology aware copy of the data (to be used for analyses where topology matters) and use the original geometry (containing circular strings) when we have to (re)generate output geometry. BTW: Is cgal/sfcgal going to replace geos eventually? It would be great if this limitation (= curved geometries being a second class citizen) was somehow lifted in future versions PostGIS. It could be, once it gets to support everything currently supported by the GEOS backend and proves to be better maintained :) --strk; ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users Huhu you are touching a sensitive point. Mostly GEOS and CGAL don't share the same purpose, so it is unfair to compare them. If you are going to need precise computation on abstract type, you want a math lib (cgal). If you are doing approximate computation because your data is not precise anyway, GEOS is fast and reliable. That being said, PostGis Topology uses neither CGAL or GEOS, and that could improve greatly performance and add functions. This is an open topic. Cheers, Rémi-C ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] curves geometries and topology
As long as they are simple circles, you should be able to convert back and forth between curves and lines. Splines, beziers, et al, are probably not going to work at all and would need to be approximated as lines. http://boundlessgeo.com/2012/01/getting-curvey/ On Mon, Apr 14, 2014 at 8:02 AM, Reijer Copier reijer.cop...@idgis.nlwrote: Dear list, According to the documentation, PostGIS Topology only supports edges made out of linestrings. However, we are currently working on a dataset that contains lots of curved geometries. It would be great if we could build topology based on those curved geometries without having to convert them to linestrings first. Is this possible? If not, what is the best way to deal with this situation? -- Kind regards, Reijer Copier IDgis bv ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users -- Eric Ladner ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
