[GRASSLIST:9000] Re: sqlite and v.rast.stats problem

[Hamish]

> The second problem is that this is horribly inefficient for my
> statistics problem. I have a huge DEM with lots of small blocks that I
> am generating statistics for. It takes several minutes for each
> calculation because the whole map is used for each call to r.univar.
> Is it possible to clip the null areas from the map region to speed up
> the calculation from r.univar?

Yes, save the region with g.region, extract the small block to a new map
with r.mapcalc, 'g.region zoom=' to the extent of the new data in
r.mapcalc, run r.univar, then restore the old region. Maybe about the
same processing time? 

maybe [reclass and] use r.stats instead on the full map?


Hamish