[GRASS5] Re: [GRASSLIST:7071] Re: Limitation in r.what

[Markus Neteler]

(moved to grass5)

On Wed, Jun 08, 2005 at 02:47:48PM +1200, Hamish wrote:
> > > I would like to query more than 150 raster layers with r.what, to
> > > extract the time evolution of a variable in a certan point of a map
> > > (each layer is for a different time step) but i do receive the
> > > error:
> > > 
> > > r.what: can only do up to 150 cell files, sorry.
> > > 
> > > The call is:
> > > 
> > > r.what input=$RASTERMAPS east_north=$myeast, $mynorth
> > > 
> > > where $RASTERMAPS is the list of my files.
> > > 
> > > How can I overcome this problem?
> ..
> > Not sure if it's the best way but maybe a bash script will do the
> > trick
> > 
> > for i in <list of raster maps>
> > do r.what input=$i  east_north=$myeast, $mynorth >> temp
> > done
> > 
> > This will (hopefully) append the value to the temp file. Just not
> > entirely sure if the redirect (>>) should be inside the loop or after
> > the done statement. Can't test it right now
> 
> or do less than 150 raster maps at a time and cat the resulting pieces
> together...
> 
> 
> grass6/raster/r.patch/nfiles.h says:
> 
> 
> /* The number of cell files that can be patched together.
>  *
>  * All cell files will be opened at one time, so this number can not
>  * be arbitrarily large.
>  *
>  * Must be smaller than MAXFILES as defined in lib/gis/G.h which 
>  * in turn must be smaller than the operating system's limit.
>  *  (Given by `cat /proc/sys/fs/file-max` in Linux 2.4)
>  */
> 
> 
> (same issue for r.series, r.patch, r.what, ...)

Wouldn't it be better to define a single value in gis.h
for all related r.* modules?
 
 
> The first line of  grass6/raster/r.what/main.c is:
> #define NFILES 150
> 
> You can make that (a bit less than) MAXFILES in lib/gis/G.h
> (currently 256) without too much worry.
> 
> Beyond that you'll need to change MAXFILES too. Probably 1024 files 
> is ok, my system says 200k is file-max. But depends on your system.
> 
> 
> Maybe I'll clean these up to reference G.h. The number can't be
> 'arbitrarily large' due to resource constraints. Search the mailing 
> lists for emails from Glynn on the subject for a better explanation.
> 
> 
> 
> Hamish

Markus