You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the LC NetCDF product 1.6.1, the attribute _FillValue=0 is used on the lccs_class variable. We currently have an unsolved issue in xarray which causes a conversion of any variable with a _FillValue into a float64. The reason is that so NaN can be used instead of _FillValue and various numpyand scipy functions become applicable that can filter out NaNs (e.g. nanmean).
According to a clarification on _FillValue, missing_value, valid_xxx, the use of valid_min=1 and valid_max=220 is sufficient and therefore the _FillValue could be simply dropped as 0 is not the only missing data / no data in the possible uint8 range.
The text was updated successfully, but these errors were encountered:
In the LC NetCDF product 1.6.1, the attribute
_FillValue=0
is used on thelccs_class
variable. We currently have an unsolved issue in xarray which causes a conversion of any variable with a_FillValue
into afloat64
. The reason is that soNaN
can be used instead of_FillValue
and variousnumpy
andscipy
functions become applicable that can filter outNaN
s (e.g. nanmean).According to a clarification on _FillValue, missing_value, valid_xxx, the use of
valid_min=1
andvalid_max=220
is sufficient and therefore the_FillValue
could be simply dropped as0
is not the only missing data / no data in the possibleuint8
range.The text was updated successfully, but these errors were encountered: