@jhamman good idea 👍

@clarkfitzg - in your eyes, is this done? If so, I'll give it another look?

@jhamman Yes, ready for review.

Will finish this up tomorrow morning.

For interactive use, I think plotting squeezed data is probably what users want. But I worry that this will make the result of plotting functions less predictable. Maybe we should restrict squeezing the data to only the magic DataArray.plot() method?

For axis titles, what happens if the data array has many scalar coordinates? Scalar coordinates can be directly assigned, not just created from indexing. My sense is that labelling facets will be more reliable if it is done explicitly in the same code that is doing the indexing/groupby.

Maybe we should restrict squeezing the data to only the magic DataArray.plot() method?

Works for me. The error message should be informative enough in the other cases.

Scalar coordinates- this is the current behavior:

In [16]: a = xray.DataArray(np.random.randn(3, 4))

In [17]: a.coords['foo'] = 123

In [18]: a.plot()

So scalar coordinates do show up in the title, which is what I was going for. I could add a test for this.

labelling facets will be more reliable if it is done explicitly in the same code that is doing the indexing/groupby.

That's the approach that I've moved to after spending more time on it. This PR adds the argument add_labels to 2d plotting functions. Now for the faceting I'll actually use add_labels=False.

OK, seems pretty reasonable on the title then. Just note that some datasets (especially from model output) contain a lot of scalar variables, so that could easily get unwieldy. We might need to revisit this once it gets more use in the wild. For example, truncating the title at some reasonable line length might be a good idea.

feel free to merge