issues: 287186057

We're getting a single test failure in the plot tests on master (link to travis failure. I haven't been able to reproduce this locally yet so I'm just going to post here to see if anyone has any ideas.

Code Sample

```python ___ TestDatetimePlot.test_datetime_line_plot _____ self = <xarray.tests.test_plot.TestDatetimePlot testMethod=test_datetime_line_plot> def test_datetime_line_plot(self): # test if line plot raises no Exception

  self.darray.plot.line()

xarray/tests/test_plot.py:1333:

xarray/plot/plot.py:328: in line return line(self._da, args, *kwargs) xarray/plot/plot.py:223: in line _ensure_plottable(x)

args = (<xarray.DataArray 'time' (time: 12)> array([datetime.datetime(2017, 1, 1, 0, 0), datetime.datetime(2017, 2, 1,... 12, 1, 0, 0)], dtype=object) Coordinates: * time (time) object 2017-01-01 2017-02-01 2017-03-01 2017-04-01 ...,) numpy_types = [<class 'numpy.floating'>, <class 'numpy.integer'>, <class 'numpy.timedelta64'>, <class 'numpy.datetime64'>] other_types = [<class 'datetime.datetime'>] x = <xarray.DataArray 'time' (time: 12)> array([datetime.datetime(2017, 1, 1, 0, 0), datetime.datetime(2017, 2, 1, ...7, 12, 1, 0, 0)], dtype=object) Coordinates: * time (time) object 2017-01-01 2017-02-01 2017-03-01 2017-04-01 ... def _ensure_plottable(*args): """ Raise exception if there is anything in args that can't be plotted on an axis. """ numpy_types = [np.floating, np.integer, np.timedelta64, np.datetime64] other_types = [datetime]

    for x in args:
        if not (_valid_numpy_subdtype(np.array(x), numpy_types)
                or _valid_other_type(np.array(x), other_types)):

          raise TypeError('Plotting requires coordinates to be numeric '
                            'or dates.')

E TypeError: Plotting requires coordinates to be numeric or dates. xarray/plot/plot.py:57: TypeError ```

Expected Output

This test was previously passing

Output of xr.show_versions()

https://travis-ci.org/pydata/xarray/jobs/326640013#L1262