issue_comments

Thanks @rpgoldman !

What do you think of that change? Feel free to discard if you disagree and we'll merge your initial suggestion.

How about we keep your sentence which is much better than my first sentence, and my second sentence? I think it's important to make the point that this function does not have side effects.

Any idea why the docs build failed?

@keewis -- Yes, that is what I ended up with: making a multi-indexed pandas data array first. But I still think it would be helpful to have this information in some of the tutorial for xarray.

Also, one thing that can happen that could be addressed in this process is a non-unique multi-index. I did this with a set of experimental data, and then .to_xarray() errored out because the multi index was non-unique, which is OK for pandas apparently, but not for xarray.

Note that this isn't necessarily pathological: This happened to me because there was oversampling of some data points in my data set.

So it would be very helpful for the tutorial to address this -- if you have multiple samples at the same point in the condition space, how do you add arbitrary indexing so that you can successfully translate from a data frame with non-unique indexing to an xarray. Is there some way to do this automatically, so we have the equivalent of lat x lon x 0, lat x lon x 1, and so on?

Even the diagnostic process is a bit of a nuisance: import numpy as np foo = filtered.index[np.where(filtered.index.duplicated(False))] foo.sort_values() so it would be nice to have a step-by-step translation guide.

@max-sixty I can see about adding other times, if someone wants to propose them, but we might have to just re-poll.

We have always had our meetings early in the morning wrt PT because of people in Europe, I believe.

Doodle poll: https://doodle.com/poll/x6quxzmyu5qxqd98

27th works for me, 25th does not. Thanks for setting this up.

I have a paid Doodle account, if you would like me to set up a poll instead of doing this here.

OK, I'm not trying to be difficult, but I think someone else who knows how RST works could fix this in 2.4 seconds, or I could bang my head on it for hours.

Would someone please just fix my syntax error and merge this?

I didn't get the cross-reference syntax right. I've tried to fix it, and am waiting for a new check.

@mathause Thank you! That never occurred to me.

I thought it was something wrong with the way I was using conjunction, but now I believe it was just a matter of operator precedence.

OK, I have tried to do as you suggested.

I looked at the directions for building the docs as you suggested, and I'm afraid it's too big an ask for me. I don't use conda, and I can't spare the time to learn it. Similarly, I know enough RST to write numpy-style docstrings, but I don't have the time to learn Sphinx. With all due respect, if xarray wants contributions from people like me, something has to be done to make this easier.

I'm just doing this because I'm contributing to Arviz, which depends on xarray. I can try to fix things if I see something wrong in the "Docs" check, but that's all I can do, I'm afraid.

@dcherian I know it's kind of overloading this issue, but if you would also LMK if there's a clean way to do compound querying, the way one can in Pandas, I will also put in something about that.

foo = data_array.where(np.logical_and(other_data_array['bar'] == 12, other_data_array['baz'] < 22))

...is all I have come up with and this gets bloated and unreadable really quickly if there are multiple conjuncts. Infix conjunction seems not to work unless I take the .values of the data array subsets, which adds its own layer of bloat.

I figure this is something that lots of people will want to do, but it seems a lot harder than pd.DataFrame.query()

OK, done. Per my comments, I don't have the ability to build and check the sphinx code, so please check before merging.

Good point: that might be a lot easier than picking up the information on the way down the stack in an exception handler.

I don't know how many call sites there are for this, so no idea how difficult that would be.

I think the Right Thing here would be to raise a specific Exception class instead of ValueError in the case of dimension mismatch. Then a caller can trap that exception, add information about the variable that cannot be translated, and re-signal.

A particular ask is "how do we get names (strings) out of the label?"

When I take the list of keys from the label, I get coordinate names, but when I access the coordinate names, I don't get the coordinate values, as I would expect. How do I get these?

I guess so. Minor preference for checking for inappropriate numpy errors like this and giving a better error message, but I can see this would be difficult.

This won't be the first bewildering error message problem from **kwargs, and it won't be the last. This is a python problem, not an xarray one, per se.

@crusaderky Thanks. Yes, that was exactly what I was trying to do.

@crusaderky I see what I did wrong. I mistakenly read the documentation page for DataArray.sum() instead of DataSet.sum(). The former does support an axis argument, and I believe it makes sense there.

Blacklisting these arguments sounds like a great idea, thanks.

@max-sixty I was trying to sum across the data variables, rather than across a dimension.

What I ended up doing was taking the data for each of the data variables and adding the resulting numpy arrays. This gave me the sum across the data variables, along the dim (each data variable is one dimensional and they share a single dimension).

I'm not sure how to suggest fixing the error message, because I'm not sure what xarray doesn't like.