[postgis-users] Postgis topolgy behaviour : is it normal?

[Rémi Cura]

Hello dear mailing list.

I'm having troubles with postgis-topology
Please see the history of this mail to see the self contained example,
which involves 4 polygons crashing the toTopoGeom function

Cheers,
Rémi-C


@ STRK :
Ok I post this to mailing list.
Thanks for your answer.
As I'm a new user to postgis-topology, I keep track of what I'm missing in
doc, and I would be glad to add some high level documentation, provided
that you check that I don't write stupid things. A simple tutorial about
postgis-topology using open data could maybe do it? I guess for somebody
used to topology concept in GIS it much be much easier to understand
postgis-topology. I find your presentation of 2011 is a good starting point.

I already have latest stable release for GEOS (3.3.8), but tomorrow I can
try to go to dev version.

I'll fill ticket tomorrow as well  ;-)

Cheers,

Rémi-C

2013/7/31 Sandro Santilli <strk at keybit.net>

> On Wed, Jul 31, 2013 at 05:13:38PM +0200, Rémi Cura wrote:
> > Hello , and thank you for your answer. I realized that mbr is "maximum
> > bouding rectangle"...
>
> "minimum"
>
> > I was guessing that you stored face geometry to avoid to compute it each
> > time, and couldn't find the right place in documentation saying
> otherwise.
>
> No, I'm not storing it.
> Up for a patch for documentation ?
>
> > I'm still trying to work with postgis-topology extension :-)
> >
> > By the way, do you prefer me to ask question directly or on the postgis
> > mailing list?
>
> Mailing list please.
>
> > I have encoutered a problem using the extension, while loading data.
> [...]
> > I get an error  : "ERROR:  Spatial exception - geometry intersects edge
> 1"
> > .
> > I don't understand why I get an error as this is a legitimate spatial
> > partitionning data. All geometry are valid according to postgis and qgis
> > They don't st_overlaps, but they do st_intersects (boundary)
> >
> > Is this because postgis topology doesn't support having areal inside
> areal?
>
> Sounds like a robustness issue. Could you please file a ticket with
> the data ? Things may change by reducing the tolerance, or using a newer
> GEOS (snap operation has been made somewhat better in current dev branch).
>
> --strk;
>
>
> >
> > Thanks a lot for your time,
> >
> > Cheers,
> >
> > Rémi-C
> >
> > 2013/7/26 Sandro Santilli <strk at keybit.net>
> >
> > > On Wed, Jul 24, 2013 at 02:01:29PM +0200, Rémi Cura wrote:
> > > > Hello,
> > > > I'm trying the postgis topology extension, and I'm running into
> > > something i
> > > > don't understand while re-enacting your example given in 2011
> > > presentation.
> > > >
> > > > First problem :
> > > > If I get it correctly, we create 3 points, then 3 edges between these
> > > > points. After this a face it automatically created.
> > > > Then when creating the 4th edge (crossing as a diagonal), the strange
> > > > behaviour happens :
> > > > it correctly creates a new face, but the 2 faces are squares, and not
> > > > triangles.
> > > >
> > > > I checked in the "face" table, and both faces are the same (and
> > > represent a
> > > > square polygon). A normal behaviour would be to have triangles.
> > > > I used the function polygonize() to update the face, with no changes.
> > > > I updated GEOS, no changes
> > >
> > > You might be confusing the "mbr" field of the face table
> > > with the face geometry. See if ST_GetFaceGeometry helps.
> > >
> > > > Second problem :
> > > > It is miscellaneous, but when deleting the postgis-topology postgres
> > > > extension, the topology schema created aren't removed (the 'conf'
> schema
> > > > using your example).
> > > > As it has no senses without the extension, I guess we could expect
> to be
> > > > deleted along.
> > >
> > > Someone actually asked to go the other way: forbid deletion if tables
> > > depend on it. So this is an open issue.
> > >
> > > > NB 0 :
> > > > the "face" table : the 2 polygon are identical
> > > > 0;""
> > > > 2;"POLYGON((10 -90,10 20,100 20,100 -90,10 -90))"
> > > > 1;"POLYGON((10 -90,10 20,100 20,100 -90,10 -90))"
> > >
> > > Right, those are the minimum bounding rectangles.
> > >
> > > --strk;
> > >
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20130731/03322d8d/attachment.html>