pandas date_range as index causes TypeError on repr

[jonathanstrong]

love this library btw, much thanks.

>>> import pandas as pd
>>> import numpy as np
>>> import xray
>>> pd.__version__
u'0.17.1'
>>> np.__version__
'1.10.4'
>>> xray.__version__
'0.6.1'
>>> import pytz
>>> import datetime
>>> datetime_index = pd.date_range(start=datetime.datetime.now(), tz=pytz.timezone('America/New_York'), periods=32, freq='1h')
>>> da = xray.DataArray(data=a, dims=['example', 'channel', 'row', 'column'], coords={'example': datetime_index, 'channel': np.arange(1), 'row': np.arange(28), 'column': np.arange(28)})
>>> print da
Traceback (most recent call last):
[...]
TypeError: data type not understood

[jonathanstrong]

sorry, forgot to mention, calling to_list() on the datetime_index allows it to work.

>>> da = xray.DataArray(data=a, dims=['example', 'channel', 'row', 'column'], coords={'example': datetime_index.tolist(), 'channel': np.arange(1), 'row': np.arange(28), 'column': np.arange(28)})
>>> print da
<xray.DataArray (example: 32, channel: 1, row: 28, column: 28)>
array([[[[ 0.2415703 ,  0.49185786,  0.76974749, ...,  0.24266482,
           0.51117921,  0.23183344],
         [ 0.5396343 ,  0.7049137 ,  0.46284154, ...,  0.14457754,
           0.5515094 ,  0.06407724],
         [ 0.6545502 ,  0.79442538,  0.61539481, ...,  0.08185602,
           0.07505339,  0.71784597],
         ..., 
         [ 0.74341592,  0.58249646,  0.9264281 , ...,  0.4105471 ,
           0.18713031,  0.38448372],
         [ 0.10294858,  0.08118397,  0.03903754, ...,  0.53996466,
           0.06814973,  0.37088457],
         [ 0.79747167,  0.71836927,  0.3760608 , ...,  0.00207118,
           0.23272135,  0.39850314]]],


       [[[ 0.54520578,  0.82793812,  0.27934314, ...,  0.47675878,
           0.35750073,  0.81324566],
         [ 0.34234289,  0.26131611,  0.1741479 , ...,  0.23433841,
           0.70296677,  0.60834615],
         [ 0.49224854,  0.35010889,  0.72760039, ...,  0.9655428 ,
           0.28976407,  0.94971166],
         ..., 
         [ 0.52062892,  0.62694481,  0.74069292, ...,  0.14967457,
           0.57194567,  0.20518411],
         [ 0.30441499,  0.29161604,  0.49102084, ...,  0.81442707,
           0.18248721,  0.88117849],
         [ 0.91370959,  0.93829229,  0.79401123, ...,  0.26210045,
           0.00614292,  0.75532125]]],


       [[[ 0.08424663,  0.51539268,  0.15021821, ...,  0.6983625 ,
           0.2479839 ,  0.45416423],
         [ 0.41073492,  0.30244808,  0.23658923, ...,  0.64526482,
           0.39558868,  0.77243522],
         [ 0.91946504,  0.14702082,  0.44866339, ...,  0.88202992,
           0.7439115 ,  0.69984251],
         ..., 
         [ 0.33138424,  0.65155252,  0.512169  , ...,  0.12812792,
           0.11201082,  0.52092288],
         [ 0.40670197,  0.04778641,  0.8280053 , ...,  0.96303147,
           0.00902775,  0.27478836],
         [ 0.63061898,  0.65237575,  0.44703296, ...,  0.26827483,
           0.62652144,  0.33875215]]],


       ..., 
       [[[ 0.02181055,  0.28960575,  0.20066007, ...,  0.96958491,
           0.48909334,  0.53350904],
         [ 0.32687157,  0.87064061,  0.75603015, ...,  0.51215617,
           0.22192232,  0.53660243],
         [ 0.53602119,  0.78761882,  0.61880669, ...,  0.40650488,
           0.21773892,  0.43152135],
         ..., 
         [ 0.6085028 ,  0.88383046,  0.72063395, ...,  0.62093148,
           0.28592564,  0.00169664],
         [ 0.47644684,  0.08029548,  0.2617365 , ...,  0.71717209,
           0.82374822,  0.67171277],
         [ 0.91905357,  0.54944082,  0.83455056, ...,  0.50262155,
           0.20663823,  0.0462887 ]]],


       [[[ 0.38780989,  0.28305091,  0.32381702, ...,  0.55012613,
           0.91371051,  0.18266697],
         [ 0.27822897,  0.19429443,  0.56619312, ...,  0.43792798,
           0.28703397,  0.48266462],
         [ 0.31136054,  0.81075836,  0.63331665, ...,  0.71262749,
           0.2267146 ,  0.81515838],
         ..., 
         [ 0.98302595,  0.85393531,  0.43698721, ...,  0.99933458,
           0.80326649,  0.87700763],
         [ 0.91663997,  0.30024646,  0.5544715 , ...,  0.5316829 ,
           0.2144375 ,  0.90622478],
         [ 0.46303225,  0.47611954,  0.17224521, ...,  0.6319952 ,
           0.22626302,  0.23822273]]],


       [[[ 0.20930888,  0.98717439,  0.07440003, ...,  0.93006893,
           0.87313465,  0.55199699],
         [ 0.56638815,  0.6957562 ,  0.14855453, ...,  0.5018839 ,
           0.88542717,  0.7561218 ],
         [ 0.46551802,  0.96133623,  0.94580378, ...,  0.09468893,
           0.30283214,  0.42860321],
         ..., 
         [ 0.14302203,  0.23090552,  0.84507949, ...,  0.89370003,
           0.23141406,  0.94566799],
         [ 0.75805896,  0.9181345 ,  0.63149841, ...,  0.18180786,
           0.94189951,  0.28630845],
         [ 0.02315291,  0.90541336,  0.86396254, ...,  0.36885415,
           0.57690491,  0.60685422]]]])
Coordinates:
  * column   (column) int64 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 ...
  * example  (example) datetime64[ns] 2016-01-12T15:31:25.870931 ...
  * channel  (channel) int64 0
  * row      (row) int64 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ...

[shoyer]

We can't handle pandas's custom dtype (e.g., datetime64[ns] with a timezone), and right now we fail very loudly when that happens.

We've encountered a few examples of this recently -- I'll see if I can clean this up for the next release.