[GRASSLIST:5375] Re: idea for new tool

[Eric G. Miller]

On Fri, Jan 24, 2003 at 08:26:52AM +1100, Gordon Keith wrote:

> Have a look at s.cellstat, 
> http://pweb.jps.net/~egm2/grass/s.cellstats/s.cellstats.html ,
> its a combination of s.univar and s.windavg with very acceptable 
> performance.
[snip]
> I'm not sure of the current status of s.cellstat. It was written by Eric 
> Miller back in September. He was going to add it to the source tree, 
> but I don't know if he has or what version.

I've just committed it to the 5.0 HEAD.  I've played around with a
couple different types of sorts to see if it made much difference in
execution time.  I didn't find much difference with my datasets, but
maybe for really large ones the change from qsort will have an impact
(up to the point where the OS has to go to swap...).  You can play with
the different sorts, if you like (see the Gmakefile).  Otherwise, it's
unchanged for months...

-- 
echo ">gra.fcw at 2ztr< eryyvZ .T pveR" | rot13 | reverse