issue_comments

At our institute we use Python (both v2.7 and 3.x) exclusively on x64 systems to make use of the entire memory on our workstations. Since all relevant (binary) scientific packages are released as 64-bit versions, 32 bit Python is obsolete for our use cases.

Just for reference (I don't develop on Windows), do we need to test both 32 and 64 bit Windows installs?

As some binary dependencies are involved, both have to be tested.

Just for reference (I don't develop on Windows), do we need to test both 32 and 64 bit Windows installs? It's not a big deal to add another configuration to the build matrix, but currently we only test Python 2.7 on 32-bit and 3.6 on 64-bit: https://github.com/pydata/xarray/blob/v0.9.1/appveyor.yml

I've just run into the same issue with the AssertionError on Windows 7 x64 with Python v2.7, netcdf4 1.2.4, numpy 1.11.3, pandas 0.19.2, xarray 0.8.2 and 0.9.1 (all packages from Anaconda repo).

Downgrading pandas to 0.18.1 solved the problem for me.

Sounds like we need to add 64-bit Python 2.7 on Windows to our build matrix!

And also a more robust check for integer types in variable.py.

@JamesSample excellent work confirming the limited scope.

I am quite sure that the issue is with this line: https://github.com/pydata/xarray/blob/v0.9.1/xarray/core/variable.py#L375-L376

dims = tuple(dim for k, dim in zip(key, self.dims) if not isinstance(k, (int, np.integer)))

When you run the test case I added in #1184, you will see that, inside __getitem__, key is a tuple with an empty array inside (key = (np.array(0, dtype=np.int64),)).

With that value, isinstance(key[0], (int, np.integer)) is False on 64 bit Windows and so dims=('time',), which has length 1.

But values = self._indexable_data[key] gives a zero-dimension array, such that values.ndim == 0...

I am unable to tell which of the two sides of the assertion expression is unexpected...

@j08lue Thanks for your response - that's interesting. I'm not familiar with Appveyor and hadn't realised what builds were being used.

As a test, I've just installed 32-bit WinPython 2.7 and xarray (still on my 64-bit Windows 10 machine) and everything works fine, so this does appear to be a specific issue with xarray and 64-bit Python 2.7.

@JamesSample Thanks for restoring my credibility a bit here... But, no, I did not figure this out yet.

Appveyor apparently does not have the combination Windows 64 bit + Python 2.7 (https://ci.appveyor.com/project/jhamman/xarray-injyf/build/1.0.659), maybe that is why it does not reproduce the error?

Hi j08lue,

Just wondering if you ever figured this out? I'm having exactly the same issue on 64-bit Windows 10, using WinPython 2.7 and the following:

xarray 0.9.1 Pandas 0.19.1 Numpy 1.11.2

If I've understood the comments here correctly, it sounds as though others are struggling to reproduce the problem - even on Windows - so presumably it's something strange with our installations?

Yes, this is Python 2. :flushed: Let's see what Travis & Co say.

Sorry for the delay getting back to you.

I'm having trouble reproducing this on OS X. I assume you're running Python 2? Maybe this is a Windows thing? I'm a little baffled here.

One thing that would be very helpful is a pull request with a test case that fails our continuous integration tests (which include testing Windows).