xarray does not seem to work with timezone aware indexes

[brunorigal]

Using the following code, the created index is not a time index but has object type:

import numpy as np
import pandas as pd
import xarray as xr

data = xr.DataArray(np.random.random(5), dims=['time'], coords={'time': pd.date_range("2000-01-01", periods=5).tz_localize('UTC')})
print(data.time)

This can severely limit the practicality of time indexes. It also means that methods such as resample won't work.

Environment:
I use xarray==0.18.2 with ubuntu

[max-sixty]

I think this a problem with the repr rather than the index:

  ...: print(data.time)
<xarray.DataArray 'time' (time: 5)>
array([946684800000000000, 946771200000000000, 946857600000000000,
       946944000000000000, 947030400000000000], dtype=object)
Coordinates:
  * time     (time) object 946684800000000000 ... 947030400000000000

In [2]: data.indexes['time']
Out[2]:
DatetimeIndex(['2000-01-01 00:00:00+00:00', '2000-01-02 00:00:00+00:00',
               '2000-01-03 00:00:00+00:00', '2000-01-04 00:00:00+00:00',
               '2000-01-05 00:00:00+00:00'],
              dtype='datetime64[ns, UTC]', name='time', freq='D')

But I wouldn't be surprised if there were issues with this sort of index — please post issues if you find them.

[brunorigal]

For example resample yields the following error:

>>> data.resample(time='1d')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/xarray/core/common.py", line 1173, in resample
    resampler = self._resample_cls(
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/xarray/core/resample.py", line 175, in __init__
    super().__init__(*args, **kwargs)
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/xarray/core/groupby.py", line 348, in __init__
    full_index, first_items = self._get_index_and_items(index, grouper)
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/xarray/core/groupby.py", line 449, in _get_index_and_items
    first_items = s.groupby(grouper).first()
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/pandas/core/series.py", line 1720, in groupby
    return SeriesGroupBy(
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/pandas/core/groupby/groupby.py", line 560, in __init__
    grouper, exclusions, obj = get_grouper(
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/pandas/core/groupby/grouper.py", line 719, in get_grouper
    binner, grouper, obj = key._get_grouper(obj, validate=False)
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/pandas/core/resample.py", line 1484, in _get_grouper
    r = self._get_resampler(obj)
  File "/home/bruno/miniconda3/envs/torch/lib/python3.8/site-packages/pandas/core/resample.py", line 1476, in _get_resampler
    raise TypeError(
TypeError: Only valid with DatetimeIndex, TimedeltaIndex or PeriodIndex, but got an instance of 'Index'

[dcherian]

The resample issue is tracked in #1490. In general, (IIUC) we use numpy arrays for the data and those don't support timezones. So it's a mess (#1490 (comment))