[Qgis-user] Multipolygons with errors ?

David J. Bakeman dbakeman at comcast.net
Mon Mar 25 10:20:27 PDT 2013


I'm bringing this back to the list so others can comment/help.

On 03/25/2013 09:22 AM, Hugues François wrote:
> The shapefile contains only the simple polygons i would like to merge to make a single (valid) multipolygon (simple ones come from an invalid  multi dumped into valid single parts).
>
> I hope this is clearer but i'm not really sure. With simple words simple geometries are valid but their union is not despite its topology seems to be ok.
OK.  So if I take your shape file and run Vector/Geometry
Tools/Singlepart to multipart.  Then run Vector/Geometry Tools/Check
Validity on the results I too get the errors you mention.  This sounds
like a bug but I'm not a GIS expert so what do others think?

Note grass doesn't report any topo errors on either the singlepart or
multipart shape files.
>
> Hugues 
>
> "David J. Bakeman" <dbakeman at comcast.net> a écrit :
>
> On 03/25/2013 08:44 AM, Hugues François wrote:
>> Hello,
>>
>> Thanks for your answer. I work with 1.8.0 lisboa on Windows.
>>
>> Did you use the same tool as i did to check geometry validity (vector/geometry tool/check validity) with a *merged* geometry from the 2 ones i sent ? 
> I did use the same tool.  I'm not sure what you mean by *merged*?  I ran
> it on the shape file I found in the rar you attached to your post.
>
> I also double checked it by loading it into grass which reported no
> topology problems.
>> Hugues 
>>
>> "David J. Bakeman" <dbakeman at comcast.net> a écrit :
>>
>> Which version of QGIS are you using?  I just tried your shape file on 1.8.0-Lisboa on Linux Fedora 18 and no geometry errors were reported.
>>
>> On 03/25/2013 07:19 AM, Francois Hugues wrote:
>>
>>
>> 	Hello,
>> 	
>> 	I received a layer with a lot of "bow tie" polygons which are invalid since OGC specifications tell us polygon's boundaries cannot cross itself. I first used st_makevalid in postgis which throw me a multipolygon geometry composed of two polygons which share one point. But if I look it in qgis and try to look for errors (vector/geometry tools/check geometry validity) i find errors like "segment 0 of ring 0 of polygon 0 intersects segment 0 of ring 0 of polygon 1 at 888252.630181, 6381825.28528".
>> 	
>> 	So I tried somethings outside postgis : i exported my data as exploded simple polygons from the original layer to a shp file (like the one attached to this message - the example of the turtle - if you would like to reproduce the problem). I tried to merge simple polygons with qgis and ran the validity checker which throw me the same errors as before.
>> 	
>> 	I definitely think there are no errors with this kind of multipolygon even if its components may share some points (here, there is only one, but we could find more complicated examples), they seem "topologically" correct to me. Am i wrong ?
>> 	
>> 	Hugues.
>> 	
>> 	--
>> 	
>> 	 
>> 	
>> 	Hugues FRANÇOIS
>> 	
>> 	hugues.francois at irstea.fr
>> 	
>> 	
>>
>> 	 
>> 	
>> 	_______________________________________________
>> 	Qgis-user mailing list
>> 	Qgis-user at lists.osgeo.org
>> 	http://lists.osgeo.org/mailman/listinfo/qgis-user
>> 	
>>
>>
>>




More information about the Qgis-user mailing list