[postgis-users] ST_MakeValid

[Stephen V. Mather]

Well, in the interest of full disclosure, as I interrogated the output from
the pgsql approach, it appears to have failed (silently?) on some of the
output-- so there are fewer input features vs. output, and in non-intuitive
ways (loss of some large polygon features.  In other words, it's faster at
the cost of being incorrect.  My apologies I didn't notice this before
posting.

I'll re-run tonight with the native function and report back on
functionality.

One feature request I could see useful would be to have the ability to
specify which topology problems to fix, and which to skip for performance
reasons. 

Best,
Steve

Stephen Mather
Geographic Information Systems (GIS) Manager
(216) 635-3243
svm at clevelandmetroparks.com
clevelandmetroparks.com





-----Original Message-----
From: Sandro Santilli [mailto:sandro.santilli at gmail.com] On Behalf Of
'Sandro Santilli'
Sent: Tuesday, April 17, 2012 4:50 AM
To: Stephen V. Mather
Cc: 'PostGIS Users Discussion'
Subject: Re: [postgis-users] ST_MakeValid

On Mon, Apr 16, 2012 at 03:05:21PM -0400, Stephen V. Mather wrote:
> > IIRC there's a difference with a multipolygon composed by two
> > overlapping rectangles.
> 
> Rereading, does ST_MakeValid also test for overlapping polygons? 

It handles those too, yes.
Basically it nodes and polygonizes all boundaries of input polygons
and then maximizes the vertices retained in output by toggling
between interior and exterior at every boundary. When impossible,
the line is retained as a linestring.

--strk; 

  ,------o-. 
  |   __/  |    Delivering high quality PostGIS 2.0 !
  |  / 2.0 |    http://strk.keybit.net - http://vizzuality.com
  `-o------'