How to get all variables with the same standard name?

[malmans2]

A few times I needed to get the name of all coordinates with the same standard name.
Did I miss a public way to do this? Otherwise, the easiest thing I could come up with is this:

cf_xarray.accessor.apply_mapper(cf_xarray.accessor._get_axis_coord, ds, standard_name)

If there isn't a public function to do this already, maybe it could be helpful. For regular variables it seems easier, as it can be done through xarray:

ds.cf[[standard_name]].data_vars

[dcherian]

You could do ds.cf[[standard_name]].coords but yes we need some API to expose all the "mappers" (#18).

Maybe

ds.cf.get_names(standard_name=None, axis=None, coordinate=None, measure=None) -> List[str]

where axis is in ["X", "Y", "Z", "T"] and coordinate is in ["latitude", "longitude", "vertical", "time"] and measure is in ["area", "volume"]

What do we do when multiple kwargs are not None; apply an OR condition, or AND?

I don't think there's a similar function in xarray/pandas/python (except filter_by_attrs): I didn't see anything similar during a quick scan through Iris and cf-python's API (can you double check?)

cc @kthyng

[dcherian]

We could also change .axes .coordinates, .standard_names, and .cell_measures to return dicts mapping "standard name" to variable name. I think I may prefer this actually.

[malmans2]

ds.cf[[standard_name]].coords would not give me just the coordinates I'm looking for, but probably I don't have the best setup in the first place.
Anyway, I also think it would be better to get dictionaries out of .axes, etc... (which to me is what describe() is kind of suggesting).
I should be able to draft a PR tomorrow!