issue_comments

@dcherian It works by netCDF4, but not for xarray: ``` file = Dataset('ds1.nc') print (file.variables['time'],'\n')

with xr.open_dataset('ds1.nc') as f: print (f.time.attrs) Output: <class 'netCDF4._netCDF4.Variable'> int64 time(time) units: hours since 2015-01-01 unlimited dimensions: current shape = (3,) filling on, default _FillValue of -9223372036854775806 used

OrderedDict() ```

What's the difference between units and attrs?

@xylar Thanks! I just found another question similar to this one.

I've tried some operations: with xr.open_dataset('merge.nc') as f: print (f['temperature'],'\n') print ('---------------------------') print (f.mean(dim='time')) print ('---------------------------') print (f['temperature'].loc[:,:,'2015-01-05T04:00:00',]) print ('---------------------------')

It works fine: ``` <xarray.DataArray 'temperature' (x: 2, y: 2, time: 6)> array([[[-0.022611, -1.428088, -0.655508, 0.977389, -0.428088, 0.344492], [ 0.430102, 0.996973, -0.882054, 1.430102, 1.996973, 0.117946]],

   [[ 0.157233, -0.230397, -0.505775,  1.157233,  0.769603,  0.494225],
    [-0.075826, -1.933904, -0.823982,  0.924174, -0.933904,  0.176018]]])

Coordinates: lon (x, y) float64 ... lat (x, y) float64 ... * time (time) datetime64[ns] 2015-01-05T04:00:00 2015-01-05T05:00:00 ... Dimensions without coordinates: x, y

<xarray.Dataset> Dimensions: (x: 2, y: 2) Coordinates: lon (x, y) float64 ... lat (x, y) float64 ... Dimensions without coordinates: x, y Data variables: temperature (x, y) float64 -0.2021 0.6817 0.307 -0.4446

<xarray.DataArray 'temperature' (x: 2, y: 2)> array([[-0.022611, 0.430102], [ 0.157233, -0.075826]]) Coordinates: lon (x, y) float64 ... lat (x, y) float64 ... time datetime64[ns] 2015-01-05T04:00:00 Dimensions without coordinates: x, y

```

@zxdawn, I was able to verify that, by adding decode_times=False to your code, I get: ``` keys of merge: odict_keys(['lon', 'lat', 'temperature', 'time'])

time of merge: <class 'netCDF4._netCDF4.Variable'> int64 time(time) units: hours since 2015-01-01 unlimited dimensions: current shape = (6,) filling on, default _FillValue of -9223372036854775806 used ```

Depending on your needs, you might be able to get away with calling open_mfdataset with decode_times=False. This should leave your time coordinate as it is (i.e. integer hours since 2015-01-01) but that will limit the types of operations you can do on the time dimension because it will not be converted to some kind of DateTime object. For example, it would be difficult to do time averaging that is aware of months, years, etc.