Following a thread I’ve opened on the grass-users list, I would like to suggest a proposal to add a flag, to the r.mapcalc command, to discrd NULL values when evaluating aggregation functions like max(), min(), mode(), etc.
Looking into the code of, for example., the max() function [1], it seems that if a NULL value is in the set of the arguments, the result is set to NULL (if I understand it correctly).
With the proposed flag, I would like to ask r.mapcalc to do not consider the NULLs, and to the calculations only on the not-NULL subset of values.
This would let us do filtering (with row/columns offsets) as it is done in r.neighbors, where the NULL values are directly discarded.
giovanni
[1] http://trac.osgeo.org/grass/browser/grass/trunk/raster/r.mapcalc/xmax.c
Hi Giovanni, all,
On Wed, Jun 20, 2012 at 12:17 PM, G. Allegri <giohappy@gmail.com> wrote:
Following a thread I’ve opened on the grass-users list, I would like to suggest a proposal to add a flag, to the r.mapcalc command, to discrd NULL values when evaluating aggregation functions like max(), min(), mode(), etc.
Looking into the code of, for example., the max() function [1], it seems that if a NULL value is in the set of the arguments, the result is set to NULL (if I understand it correctly).
With the proposed flag, I would like to ask r.mapcalc to do not consider the NULLs, and to the calculations only on the not-NULL subset of values.
Shouldn’t it be the default behavior?
ciao madi
This would let us do filtering (with row/columns offsets) as it is done in r.neighbors, where the NULL values are directly discarded.
giovanni
[1] http://trac.osgeo.org/grass/browser/grass/trunk/raster/r.mapcalc/xmax.c
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev
–
Dr. Margherita Di Leo
Shouldn’t it be the default behavior?
I supposed it too, but it isn’t 
giovanni
ciao madi
This would let us do filtering (with row/columns offsets) as it is done in r.neighbors, where the NULL values are directly discarded.
giovanni
[1] http://trac.osgeo.org/grass/browser/grass/trunk/raster/r.mapcalc/xmax.c
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev
–
Dr. Margherita Di Leo
G. Allegri wrote:
Following a thread I've opened on the grass-users list, I would like to
suggest a proposal to add a flag, to the r.mapcalc command, to discrd
NULL values when evaluating aggregation functions like max(), min(),
mode(), etc.
Looking into the code of, for example., the max() function [1], it seems
that if a NULL value is in the set of the arguments, the result is set to
NULL (if I understand it correctly).
With the proposed flag, I would like to ask r.mapcalc to do not consider
the NULLs, and to the calculations only on the not-NULL subset of values.
This would let us do filtering (with row/columns offsets) as it is done in
r.neighbors, where the NULL values are directly discarded.
In 7.0, I've added nmin, nmax, nmedian and nmode functions which
ignore nulls.
--
Glynn Clements <glynn@gclements.plus.com>