add test for 1D object array indexing

[davidtrem]

Test for issue #2414

closes #2414

[shoyer]

Maybe add:

expected_array = np.empty((), dtype=object)
expected_array[()] = np.arange(3)
expected = DataArray(expected_array)
assert_identical(expected, testarray[0])

[davidtrem]

Well...
I guess you mean:

expected_array = np.arange(3)
expected = DataArray(expected_array)
assert_identical(expected, testarray[0])

[shoyer]

expected_array = np.arange(3) gives you an array with shape (3,) and int dtype. You want a scalar NumPy with shape () and object dtype.

[davidtrem]

OK, I get you point :-)
Main issue I still have then is that I'm not able to get the 'np.arange(3) array'
from the 'expected'...
because 'expected.data' is king of empty :-(
expected.data.shape == () therefore expected.data[0] raise an IndexError
How would you access the content of 'expected' ?

[shoyer]

You could pull it out with expected.data.item()

…

On Fri, Sep 14, 2018 at 12:43 AM David Trémouilles ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In xarray/tests/test_indexing.py <#2415 (comment)>: > @@ -622,3 +622,8 @@ def test_create_mask_error(): def test_posify_mask_subindexer(indices, expected): actual = indexing._posify_mask_subindexer(indices) np.testing.assert_array_equal(expected, actual) + + +def test_indexing_1d_object_array(): + testarray = DataArray(np.array((np.arange(3), np.arange(6)))) + testarray[0] OK, I get you point :-) Main issue I still have then is that I'm not able to get the 'np.arange(3) array' from the 'expected'... because 'expected.data' is king of empty :-( expected.data.shape == () therefore expected.data[0] raise an IndexError How would you access the content of 'expected' ? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2415 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABKS1uZNHXq96w2Dbt50Bw2uBHVxy00-ks5ua14TgaJpZM4Wnb0S> .

[keewis]

if we still want this test, it should be ready for review / merging

[max-sixty]

Let's do it! Thanks for resuscitating @keewis

[dcherian]

Thanks @keewis