> Won't that only yield the same answer if you can guarantee no overlapping > geometries in the fazendas table? My understanding was that the other > table was the raster table (which it is reasonable to assume has no > overlaps), but we don't know if that's the case with this one.
Right. But you generally have to have a minimal knowledge about the topology of a layer when you query it. Querying a layer with overlaps is never the same as querying a layer without. So if you know your fazendas layer has overlaps it is a good idea to use ST_Union but that means you are getting stats about the global farming area not about each farming area. What about areas overlapping? Should they count twice? Do these overlaps make sense in reality? Maybe you have to fix then first... Pierre _______________________________________________ postgis-users mailing list [email protected] http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
