[GRASS-user] areas, isles and centroids - topology doubts about category assignment

G. Allegri giohappy at gmail.com
Fri Aug 8 12:53:01 EDT 2008


Here I am again...
I'm facing a topological problem, surely due to my ignorance.
I've dissolved a polygonal layer, composed by polygons sharing
boundaries and isolated polygons, with various holes within each
polygon. I created a foo column and have assigned value for all the
features, to be able to dissolve every boundary between adjacent
areas.
The result has the following topology composition:

nodes=386
points=0
lines=0
boundaries=375
centroids=11
areas=21
islands=21
faces=0
kernels=0
primitives=386
map3d=0

type       count        min        max
point          0          0          0
line           0          0          0
boundary       0          0          0
centroid      11          1          1
area           0          0          0
all           11          1          1

First question:
the topology dump tells me that I have 21 areas and 21 isles, bu the
"true" areas are 11. I've read with attention the dump, and I can see
that the holes in the original polygons now are considered both as
areas (with no centroid) and isles... Shouldn't be considered only an
island or an area depending on the lines "signs"?
Anyway, the problem is that if now I build the topology on this
dissolved layer, I get isles "transformed" to areas getting centroids
attached to them. That doesn't sound me conceptually correct. Could
you explain me why?

Second question:
if I use v.category to differentiate the categories (using "sum"
option) and select only "centroids", the result remains unchanged
respect to the original. Things change if I select also "area". In
this case... the categories get changed, but now I have 21 centroids:
not only the effective areas, but also isles/areas receive centroids!
So, my question is: I just want my 11 "true" areas have categories
from 1 to 11, and let the isles remain isles... How to do it?

(Third question):
would it be possible, in the future of grass, not having to produce a
second output vector when updating categories, dissolving boundaries,
etc.? It would simply mean that the original get overwritten by the
"new" one...

Thanks, and sorry for the long questions!
Giovanni


More information about the grass-user mailing list