[GRASS-dev] [GRASS GIS] #2497: Incompatile PostGIS Topology export

#2497: Incompatile PostGIS Topology export
-------------------------+--------------------------------------------------
Reporter: strk | Owner: grass-dev@…
     Type: defect | Status: new
Priority: normal | Milestone: 7.0.0
Component: Vector | Version: svn-trunk
Keywords: | Platform: Unspecified
      Cpu: Unspecified |
-------------------------+--------------------------------------------------
I've tried exporting topology to PostGIS with v.out.postgis -l and the
result was not correct.

First there was no record in TOPOSCHEMA.relation so that all the literal
TopoGeometry values (written in the output feature layer) result empty.

Second all the calls to ST_GetFaceGeometry() return a null value, which
I've yet to inspect/understand. All I noticed so far was that there were
many negative face_id, which don't really have any special meaning in
PostGIS/ISO (there's no difference between an area or an island, you only
have faces, with roles being dependent on the upper abstraction level of
the TopoGeometry).

Unfortunately, topology.ValidateTopology doesn't notice any of these
problems (this would be a PostGIS bug).

--
Ticket URL: <https://trac.osgeo.org/grass/ticket/2497&gt;
GRASS GIS <http://grass.osgeo.org>

#2497: Incompatile PostGIS Topology export
-------------------------+--------------------------------------------------
Reporter: strk | Owner: grass-dev@…
     Type: defect | Status: new
Priority: normal | Milestone: 7.0.0
Component: Vector | Version: svn-trunk
Keywords: | Platform: Unspecified
      Cpu: Unspecified |
-------------------------+--------------------------------------------------

Comment(by strk):

I've found that ST_GetFaceGeometry only returns a NULL value for faces
with a negative id (which would make sense, assuming no edge would ever
point to them as being on its right or left side.

Should these negative faces be put in a grass-specific table rather than
polluting the ISO one ?

--
Ticket URL: <https://trac.osgeo.org/grass/ticket/2497#comment:1&gt;
GRASS GIS <http://grass.osgeo.org>