Re: [OSM-dev] osm2pgsql polygon handling semantic?

Paul Norman wrote: > For the osm2pgsql pgsql backend, a polygon is formed when a way is > closed and the tag transform sets a polygon flag. If either of these is > false, it is a linear feature. ... which is then inserted into planet_osm_line, right? For some reason I would

Re: [OSM-dev] osm2pgsql polygon handling semantic?

Simon Poole wrote: > Roundabouts? Or in fact any other way that is closed but clearly a > linear feature. I think those are usually not defined as a polygon anyway at least without area=yes. So far I have found two types of tags (waterway and aeroway) which will end up in

Re: [OSM-dev] osm2pgsql polygon handling semantic?

Paul Norman wrote: > Assuming it matches some tag with a column, yes. In a case where it has > only a key with polygon,nocolumn (or phstore), it wouldn't make it into > the line table or polygon table. This does not seem to be true. Artificial waterway.osm --cut--

[OSM-dev] osm2pgsql polygon handling semantic?

Hello, looking at the code of osm2pgsql and checking with a hand-crafted .osm file I came to the conclusion that polygons in osm2pgsql are handled in the following way: If a polygon is defined in the .style file a function is called which generates the simple-feature geometry object. If this