[GRASS-dev] wierd results with r.series meth=perc90,q3

[Hamish]

Hi,

I'm getting some strange results with r.series method=perc90 and quart3.

8 input maps, partially overlapping, each map has nodata areas of NULL.
(CELL maps, 0->254 + NULL)


the perc90 results in a blank map (r.info: min=nan, max=nan) while the
quart3 map only seems to have data in slivers where some of the layers
overlap.
?

also, it would be nice if for stats like min,max, if only CELL maps are
given for input, a CELL map will be written as output. Probably the
sanest way to do this is to add an outtype=CELL option to the module?
(e.g. r.in.xyz)  I guess we should use rounding in the case of a CELL
map?



thanks,
Hamish