[postgis-users] Topology and Triggers

Bruce Rindahl rindahl at lrcwe.com
Thu May 15 11:43:27 PDT 2008


Has anyone thought about the Arc/Info coverage model?  The idea would be 
to have a table with line geometries and attributes reflecting the left 
and right polygon ids.  My thought is you do your edits on the line 
geometry table and then polygon topology is built in a separate 
(dependent, inherited??) table based on the line table using well 
defined procedures.  The only function I  don't see offhand that is not 
there is a split polygon with line.  Just my thoughts.

Bruce Rindahl

Webb Sprague wrote:
> On Thu, May 15, 2008 at 5:20 AM, George Silva <georger.silva at gmail.com> wrote:
>   
>> Hello all again,
>>
>> I´m curious. I´ve said in qgis list and some other 'desktop' GIS software
>> lists that the biggest feature missing for OSGIS is something like the
>> 'auto-complete' polygon in ArcGIS,, that would make life much easier.
>>     
>
> Indeed!
>
>   
>> I´ve been reading some things, and i´ve seem people mentioning trigger based
>> topology (eg cut out the overlaping area of digitized polygons, or,maintain
>> only the overlaping areas of the digitized polygons, and other related
>> issues).
>>     
>
> I think the hard part is deciding how you want to do this exactly,
> since all approaches involve changing the data somehow; a lot of
> people are chary of automatically deleting gaps, for example, lest
> they aren't really gaps (consider the Vatican Republic in  a map of
> euro nations...).
>
>   
>> I imagine that would be a trigger using something like after insert update
>> the_geom.
>>
>> Can you guys with more experience give me a example, so i can build my own?
>> Best pratices? Some light around this subject?
>>     
>
> Have you ever written a non-geo trigger in postgresql?  If not, I
> would practice  using the docs; I am pretty sure you could just modify
> NEW with a geo operator/ function.
>
> If so, I think a larger scale project needs to be embarked on.  The
> work so far seems to me to be quite low level without a grand plan,
> and I think a grand plan is needed.  I would start with a bunch of
> functions for modifying a geometry based on its neighbors and some
> topology constraints, then maybe a schema to store topology
> information, and only then start implementing triggers.  But this is
> armchair quarterbacking--hopefully someone with some real experience
> will say something...
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>
>
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20080515/e3046c3a/attachment.html>


More information about the postgis-users mailing list