Add an ```only_variables``` or similar option to ```xarray.open_dataset``` and ```xarray.open_mfdataset```

[jerabaul29]

Is your feature request related to a problem?

Sometimes, a variable in a nc file is corrupted or not "xarray friendly" and crashes opening a file (see for example #8072 ; I solved this on my machine by just drop_variablesing the problematic variables in practice), or reading and parsing the full file or mf-file may be expensive and time consuming, while only a couple of variables are needed.

Describe the solution you'd like

We already can exclude variables with the drop_variables arg to open_dataset (note: this is not present for now in open_mfdataset, should it be added there?), but could we also instead of saying "read all the variables instead of this list", be able to say "read only these variables"? In most case, this would be equivalent of using drop_variables=list(set(all_vars)-set(list_insteresting_vars), but in case some (many vars) may be corrupted, just getting the file opened to list these all_vars may be problematic.

Describe alternatives you've considered

drop_variables=list(set(all_vars)-set(list_insteresting_vars), but this is a lot more verbose.

Additional context

No response

[TomNicholas]

drop_variables arg to open_dataset (note: this is not present for now in open_mfdataset)

open_mfdataset accepts arbitrary kwargs that are then fed down to open_dataset. Have you tried just passing the drop_variables kwarg to open_mfdataset?

[jerabaul29]

drop_variables arg to open_dataset (note: this is not present for now in open_mfdataset)

open_mfdataset accepts arbitrary kwargs that are then fed down to open_dataset. Have you tried just passing the drop_variables kwarg to open_mfdataset?

Thanks for the recommendation! I have just tried using open_mfdataset with the drop_variables argument, and this works. So this answers this part :) . Wonder if this could be made clearer in the documentation of the open_mfdataset function somehow? :)

Guess this answers this "secondary point" of the issue, but that the issue should remain open, as this is separate from the "main point" asked :) . (just a note to people who may read the thread a bit fast :) ).

[TomNicholas]

Wonder if this could be made clearer in the documentation of the open_mfdataset function somehow? :)

@jerabaul29 would you be interested to submit a small PR which adds an example to the docstring of open_mfdataset? :)

[jerabaul29]

Wonder if this could be made clearer in the documentation of the open_mfdataset function somehow? :)

@jerabaul29 would you be interested to submit a small PR which adds an example to the docstring of open_mfdataset? :)

I will happily do it. I can look into this tomorrow or over the weekend (depending on when I have time). I tell you if I have any difficulties. If I am too overbooked and I forgot, feel free to ping me next week :) .

[jerabaul29]

@TomNicholas see #8083 regarding the small doc update :) . (if I understand well, the doc is automatically generated from the docstring in the source, right? :) ).

[TomNicholas]

the doc is automatically generated from the docstring in the source

Yes, exactly!

[jerabaul29]

Then I think that the example is now correctly added. (note that we still should discuss if only_variables is interesting / useful / worth it to add :) ).