Please note that the to_netcdf method is not properly implemented, since it is saving the original netcdf data and ignoring the actual fieldlist structure.

Please note that the to_netcdf method is not properly implemented, since it is saving the original netcdf data and ignoring the actual fieldlist structure.

I think this is what we want, right? If we are just saving the object we want the original file where possible with minimal modification from EK.

Actually I was a little unsure whether we should even be using xarray for opening and then writing, but I think that change can be part of another issue and PR.

Please note that the to_netcdf method is not properly implemented, since it is saving the original netcdf data and ignoring the actual fieldlist structure.

I think this is what we want, right? If we are just saving the object we want the original file where possible with minimal modification from EK.

Actually I was a little unsure whether we should even be using xarray for opening and then writing, but I think that change can be part of another issue and PR.

Please consider these examples:

# ds is a fieldlist with e.g. 100 fields
ds = earthkit.data.from_source("file", "my_data.nc")
f = ds[2:10]
f.save("my_file.nc")

# ds1 is a fieldlist with 10 fields
ds1 = earthkit.data.from_source("file", "my_data_1.nc")
# ds2 is a fieldlist with 20 fields
ds2 = earthkit.data.from_source("file", "my_data_2.nc")
# concatenation
ds = ds1 + ds2
f = ds[8:12]
f.save("my_file.nc")

Please note that the to_netcdf method is not properly implemented, since it is saving the original netcdf data and ignoring the actual fieldlist structure.

I think this is what we want, right? If we are just saving the object we want the original file where possible with minimal modification from EK.
Actually I was a little unsure whether we should even be using xarray for opening and then writing, but I think that change can be part of another issue and PR.

Please consider these examples:

# ds is a fieldlist with e.g. 100 fields
ds = earthkit.data.from_source("file", "my_data.nc")
f = ds[2:10]
f.save("my_file.nc")

# ds1 is a fieldlist with 10 fields
ds1 = earthkit.data.from_source("file", "my_data_1.nc")
# ds2 is a fieldlist with 20 fields
ds2 = earthkit.data.from_source("file", "my_data_2.nc")
# concatenation
ds = ds1 + ds2
f = ds[8:12]
f.save("my_file.nc")

I added a not implemented error to the NetCDFMaskedFieldList as this was producing an error.

The MultiFile looks to work if the 2 data objects are concatable.

The MultiFile looks to work if the 2 data objects are concatable.

Should be tested. But please consider this code:

# ds1 is a fieldlist with 10 fields
ds1 = earthkit.data.from_source("file", "my_data_1.nc")
# ds2 is a fieldlist with 20 fields
ds2 = earthkit.data.from_source("file", "my_data_2.nc")

f1 = ds1[2:5]
f2 = ds2[8:14]
f = f1 + f2
f.save("my_file.nc")

The MultiFile looks to work if the 2 data objects are concatable.

Should be tested. But please consider this code:

# ds1 is a fieldlist with 10 fields
ds1 = earthkit.data.from_source("file", "my_data_1.nc")
# ds2 is a fieldlist with 20 fields
ds2 = earthkit.data.from_source("file", "my_data_2.nc")

f1 = ds1[2:5]
f2 = ds2[8:14]
f = f1 + f2
f.save("my_file.nc")

But this is a problem with the to_xarray method of the NetCDFMaskedFieldList object, which I don't think is the responsibility of this issue and PR. I have added another not implemented exception for when the objects do not have a path attribute.

Just to be clear, the following currently fails, and I believe is outside the scope of this PR:

# ds1 is a fieldlist with 10 fields
ds1 = earthkit.data.from_source("file", "my_data_1.nc")
# ds2 is a fieldlist with 20 fields
ds2 = earthkit.data.from_source("file", "my_data_2.nc")
 
f1 = ds1[2:5]
f2 = ds2[8:14]
f = f1 + f2
# This will fail:
f.to_xarray()
# This will also fail:
f.to_netcdf("my_file.nc")

I have moved the "Not implemented" Exception to the correct layer, i.e. on the to_xarray method of the NetCDFMaskFieldList

hey @sandorkertesz, thanks for finishing this off :)