On 27/09/2010 20:29, Markus Metz wrote:
Micha Silver wrote:
Cheers.
Although I think I misled you a bit. I was mistaken about the orange line
color in the digitizer. It simply indicates a boundary that is wholly shared
by two adjacent areas. *Not* a topological error.
In any case, importing polygon shapefiles first as lines (not boundaries),
doing the topology cleanup, and then convert to boundaries and add centroids
seems to be the smoother way to go.
I object because
1. you would need to replicate the cleaning steps done by v.in.ogr
which has, amongst others, a snapping option
I never noticed that option.
2. adding centroids will add centroids to all areas, also those that
where holes in polygons provided by OGR, e.g. a waterbodies shapefile
with lakes and islands in lakes: the islands are not waterbodies and
should not get a centroid
I think Bryan's original problem was areas being imported without
centroids. But in the general case, I see that importing lines, then
converting to boundaries and adding centroids would be wrong.
3. you will loose attributes because there is no (easy) way to link
any attributes coming with the shapefile to newly generated centroids,
e.g. land cover/land use shapefiles
Yes, this was clear.
Markus M
Still, importing unclean shapefile polygons, and correcting the topology
is often accompanied by much hair pulling...
Thanks for the clarifications!
--
Micha
_______________________________________________
grass-user mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/grass-user
