I experience some difficulties related to v.patch and category / attributes
when merging vector maps together.
I've read the following thread :
osgeo-org.1560.x6.nabble.com/Appending-a-map-to-another-td5176882.html and
proceeded the way Markus dis explained it. FI 'm working with grass70.
*Point I : *
As Tommaso said (in mentioned thread), I noticed that the /v.category
option=sum to change the categories / doesn't seem to be needed because
grass reclass it automatically. Is this correct ?
*Point II*
More problematic, opening attributes table, I can see the following cats :
10003 30006 60009 70013 90017 120021
And /v.category input=poly_iso option=print layer=1 id=1-50/ will return
same values
but displaying it on the map shows both boundaries and areas - And
retrieving attributes on click will only work on areas. On the image below,
I have 3 areas (sometimes it is 2).
On Fri, Jun 5, 2015 at 9:22 AM, simogeo <simon.georget@gmail.com> wrote:
For information, it seems that creating not one but 3 layers make this work
correctly :
...
This sound like a bug, no ? I assume that because some polygons have same
centroid, GRASS get lost. Could it be the cause ?
Should I report it as bug or not ?
While I am not sure if this a bug, please open a ticket for it along
with an example (ideally for the North Carolina dataset). If related
sample data need to be used in your example, you may export them with
v.pack and attach to the ticket.
Thanks for your reply. Actually,I investigated a bit more and I realized that this is probably because of GRASS topological model.
As you can see, some of my generated polygons overlaps others - and some centroids may be contained into others polygons (on same vector map). I guess, GRASS doesn’t like much this situation (could you confirm ?).
Anyway, any precision would be appreciated to understand better.
If you still think it could be a bug. I’ll find time for reporting it. Thanks again.
···
On Sun, Jun 7, 2015 at 10:38 PM, Markus Neteler [via OSGeo.org] <[hidden email]> wrote:
On Fri, Jun 5, 2015 at 9:22 AM, simogeo <[hidden email]> wrote:
For information, it seems that creating not one but 3 layers make this work
correctly :
…
This sound like a bug, no ? I assume that because some polygons have same
centroid, GRASS get lost. Could it be the cause ?
Should I report it as bug or not ?
While I am not sure if this a bug, please open a ticket for it along
with an example (ideally for the North Carolina dataset). If related
sample data need to be used in your example, you may export them with
v.pack and attach to the ticket.
On Mon, Jun 8, 2015 at 10:48 AM, simogeo <simon.georget@gmail.com> wrote:
Hi Markus,
Thanks for your reply. Actually,I investigated a bit more and I realized
that this is probably because of GRASS topological model.
As you can see, some of my generated polygons overlaps others - and some
centroids may be contained into others polygons (on same vector map). I
guess, GRASS doesn't like much this situation (could you confirm ?).
Indeed, they are like towers of Hanoi....
The given link below related to Vector topology is explicit enough :
/>> Boundaries should not cross each other (i.e., boundaries which would
cross must be split at their intersection to form distict boundaries)/
Still, the behavior is a bit strange since on the (top) right of the image,
2 polygons appear to be well-formed - and it should not, as polygons on the
(bottom) right side.
I hope I'll have some more opportunities to run my script on other data,
I'll be watchful !
Thanks again.
