[postgis-devel] ST_Perimeter and ST_Perimeter3D what should Polyhedral Surface return?

Nicklas Avén nicklas.aven at jordogskog.no
Fri Aug 13 10:43:36 PDT 2010


Regina and Olivier

This about perimeter in 3D. 
When I thought about it it lead me to the question when a in polyhedral
surface becomes a polyhedron. Like a closed linestring, can be a
linestring without turning to a polygon.

The two questions connect in that, when the surface turns to a volum I
guess there is no perimeter like there is no length of line when the
linestring becomes a polygon.

Olivier, I thought about it also because of the discussion the other day
about testing if a point is inside a closed polyhedral when doing
distance-measures. If there is no way to show that it is a polyhedron
there should be the distance to the surfaces even if the point is
inside?
I mean the distance from a point to a closed linestring is the distance
to the linestring even if the point is inside the closed line.

/Nicklas






On Thu, 2010-08-12 at 20:47 -0400, Paragon Corporation wrote:
> I revised the garden test to log timings to a table postgis_garden_log.
> Mostly I just wanted to see which functions Polyhedral surface successfully
> completes.
> 
> If a step is completed without rolling back, the log_end is filled in.  I
> figure this will be later useful for timings as well.
> 
> There are a surprising number that polyhedral surfaces return answers to.
> 
> One is ST_Perimeter and ST_Perimeter3D.  This just seems to be a sum of the
> polygon perimeters.  
> 
> To me this answer is wrong since I don't think the shared edges should be
> included.
> 
> Also I noticed for Polyhedral surface, it returns an invalid multipolygon
> since edges intersect.  Do we care that ST_CollectionExtra can return
> invalid multipolygons form a valid collection of polygons?
> 
> Thanks,
> Regina
> 
> 
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
> 





More information about the postgis-devel mailing list