[postgis-users] A question on ST_RemEdgeModFace usability

Andrea Peri aperi2007 at gmail.com
Thu Apr 12 12:00:41 PDT 2012


Hi,

just some samples to start the question:

suppose to have 2 topogeometry T1 and T2 where
T1 = face1 (F1) + face2 (F2)

And T2 = face3 (F3)

and outside the universal-face

      ----------------
    /          |       \  (universe Face)
    |          |       |
    |      F1  | F2    |
    |          |       |
    |          e1      e2
    |          |       |
     \----------------/
    /                 \
    |                 |
    |        F3       |
    |                 |
    |                 |
    |                 |
     \---------------/

The edge e1 is invisible for the Topogeometry1 (T1)
so actually the postgis topology allow to remove it with the
ST_RemEdgeModFace().
And having so this new situation:

      ----------------
    /                 \
    |                 |
    |      F1         |
    |                 |
    |                 e2
    |                 |
     \----------------/
    /                 \
    |                 |
    |        F3       |
    |                 |
    |                 |
    |                 |
     \---------------/

where the TopoGeometry are:
T1 = face1 (F1)
T2 = face3 (F3)

But if the edge I need to remove is the e2 edge the situation change:
please notice the edge e2 has the universe-face on a side.

      ----------------
    /          |       \
    |          |       |   (universe face)
    |      F1  | F2    |
    |          |       e2
    |         e1       |
    |          |       |
     \----------------/
    /                 \
    |                 |
    |        F3       |
    |                 |
    |                 |
    |                 |
     \---------------/

I agree the new situation should be this:

      ----------
    /          |
    |          |          (universe face)
    |      F1  |
    |          |
    |         e1
    |          |
     \----------------
    /                 \
    |                 |
    |        F3       |
    |                 |
    |                 |
    |                 |
     \---------------/

But as I understand this seem to be impossible to have because actually
postgis topology don't allow to remove the edge e2.
Becasue this mean change the form of the topogeometry T1

Please notice this is don't compliant with the ISO specs, but in effect the
ISO specs don't know the TopoGeometry layer.

I guess this use-case was not compliant with iso specs, but it is again a
perhaps unfrequent case.

But there also this other case more frequetnly that could reduce many the
useful of ST_RemEdgeModFace:

Starting from this situation:


      ----------------
    /          |       \  (universe Face)
    |          |       |
    |          |       |
    |----------|-------|
    |    |     |       |
    | F5 e4 F4 e5      |
    |    |     |       |
    |----------|       |
    |          |       |
    |      F1  | F2    |
    |          |       |
    |         e1       |
    |          |       |
     \---------------/
    /                 \
    |                 |
    |        F3       |
    |                 |
    |                 |
    |                 |
     \---------------/

where
Topogeometry T3 = face4 (F4)

I could need to remove the edge e4, where the e4 is and edge that don't
have the universe-face on any side.

Is the edge e4 removable with the ST_RemEdgeModFace and have this result


      ----------------
    /          |       \  (universe Face)
    |          |       |
    |          |       |
    |----------|-------|
    |          |       |
    | F4       e5      |
    |          |       |
    |----------|       |
    |          |       |
    |      F1  | F2    |
    |          |       |
    |         e1       |
    |          |       |
     \---------------/
    /                 \
    |                 |
    |        F3       |
    |                 |
    |                 |
    |                 |
     \---------------/

Is this allowed ?

Thanks,

Andrea Peri


-- 
-----------------
Andrea Peri
. . . . . . . . .
qwerty àèìòù
-----------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20120412/c8f00a36/attachment.html>


More information about the postgis-users mailing list