[GRASS-user] v.net.path problem (not the shortest route)

[Markus Metz]

On Fri, Aug 19, 2011 at 11:16 AM, Jorge de Jesus <jmdj at pml.ac.uk> wrote:
>  Hi to all
>
> I've a question concerning v.net.path. I am probably missing something
> obvious :-[
>
> In the v.net.path example that I've posted yesterday
> (http://rsg.pml.ac.uk/wps/example/index.html), if we put the car over
> madrid and the destination in the node below Santa Maria (below Porto
> city), the v.net.path uses a different route going thru Vigo. Therefore
> picking a route that isnt the shortest.
>
> A screenshot of the problem can be seen here: http://derp.co.uk/6de36
>
> Basically it seems that v.net.path starts to pick the shortest path from
> the 1st node to the 2nd node independently of the total route length
> (therefore not checking if it is actually the shortest total path).
>
> Any comments and suggestions ?!

Update:

One intersection is missing: run

v.clean in=graph out=graph_clean tool=break,rmdupl

rebuild network graph from vectors graph_clean and nodes, calculate
shortest path. That should give you the real shortest path.

Markus M