[GRASS-user] computation in a location based on infos from another location

[Patrice Dumas]

On Thu, Nov 19, 2009 at 04:30:55PM +0100, Achim Kisseler wrote:
> Hi,
> 
> > use r.to.vect to map every cell to a polygon, project it to the lambert 
> > map, do the calculation on the polygon, reproject it back to the 
> 
> just run v.to.db -p option=area map=...
> 
> then you get the size for every lines cell.
> 
> To get the raster-map:
> r.mapcalc "ns_col=if(col()==1 , row() , null())"

My idea was to do the other way around. In the lonlat location (cru_maps), 
I do a polygon that spans the whole region and delimitates cells:

rows_nr=`g.region -g |grep '^rows=' |sed 's/^rows=//'`
r.mapcalc "grid_map=col() + ((row()-1) * $rows_nr)"
r.to.vect input=grid_map out=grid_map_v feature=area

Then in the Lambert map, I project and use v.rast.stats:

v.proj input="grid_map_v" location="cru_maps"
v.rast.stats vector=grid_map_v raster=lambert_map colprefix=stat

then, in the lonlat location I project back
v.proj input="grid_map_v" location="network_af" output=result_v
v.to.rast input=result_v output=result_rast column=stat_sum

Right now I am stuck at the first projection since it triggered an 
assertion...

--
Pat