ex06_missing_values
This example demonstrates missing-value handling for text and NetCDF inputs.
The text primary contains NaN, while the text secondary uses -99999 and
declares it with GMT’s -di-99999 option.
The equivalent NetCDF primary declares _FillValue. The NetCDF secondary
does not declare its sentinel, so the selector supplies it explicitly:
primary.nc?vp secondary.nc?p+n-99999 0/10 cosine 0.2
Interpolation is performed independently within each contiguous data run and
does not bridge a missing interval. Where only the primary or secondary is
available, that value is returned regardless of the merging weight. Where
neither is available, the output is NaN.
Appending +g to the interpolation option bridges every internal missing
interval before merging:
gmt merge1d inputs.merge -T0/10/0.05 -Sl+g -Fvalue -W -Gmerged.txt
The final two panels compare the interpolated primary and secondary series
with the resulting merge. Missing values outside the first and last available
samples would remain missing because +g does not extrapolate.
Run:
./ex06_missing_values.sh
The script verifies that the text and NetCDF routes give numerically
equivalent results both with and without gap interpolation, checks the
missing-data cases, and creates ex06_missing_values.png and
ex06_missing_values.pdf.
Figures
