cdffill − interpolate across bad data segments
cdffill dimension
Cdffill examines each field in each variable slice for bad data. If the bad data are sandwiched between good data in the specified dimension, the bad data are replaced by values obtained by linear interpolation from bounding good data values. Bad data at the upper or lower ends are not bounded by good data, and are therefore not replaced by interpolated values.
kestrel%cdffill height < infile > outfile
In this example all fields with the dimension "height" have bad data segments replaced by values linearly interpolated in height.
Currently, if the index field associated with the specified dimension has unevenly spaced values, the interpolation will be incorrect, as even spacing is assumed.