[GRASS-dev] r.univar: 'nan' results

[Maciej Sieczka]

Hi,

Using current 6.3 CVS I had a strange issue with r.univar. For one of
my rasters most statistics are "nan" in r.univar, which I don't
understand why and while the r.univar.sh output is OK for the same
raster.  Both outputs you'll find below. The raster is an product of
v.to.rast use=dir.

$ r.univar cieki10_2zl_dir
 100%
total null and non-null cells: 378972
total null cells: 374849

Of the non-null cells:
----------------------
n: 4123
minimum: 0
maximum: 358.264
range: 358.264
mean: nan
mean of absolute values: nan
standard deviation: nan
variance: nan
variation coefficient: nan %
sum: nan



$ r.univar.sh cieki10_2zl_dir
WARNING: This module is superseded and will be removed in future
versions of GRASS. Use the much faster r.univar instead.

Calculation for map cieki10_2zl_dir (ignoring NULL cells)...
Reading raster map...
r.stats:  100%
Calculating statistics...

Number of cells (excluding NULL cells): 4123
Minimum: 0
Maximum: nan
Range: 0
Arithmetic mean: 160.434
Arithmetic mean of absolute values: 160.434
Variance: 11639.4
Standard deviation: 107.886
Variation coefficient: 67.2463 %



What's wrong?



I couldn't reproduce this in spearfish. Below's is the r.info output.
If my raster is needed for testing I can send it.

datatype=DCELL

min=0.000000
max=358.264295

nsres=10
ewres=10

north=5684100
south=5676270
east=601480
west=596640

Data Source:
   Vector Map: cieki10_2zl in mapset melio2
   Original Scale from Vector Map: 1:1
Data Description:
   generated by v.to.rast
Comments:
   v.to.rast input="cieki10_2zl" output="cieki10_2zl_dir" use="dir" lay\
   er=1 value=1 rows=4096

Maciek