issues: 268725471

This data as json

id node_id number title user state locked assignee milestone comments created_at updated_at closed_at author_association active_lock_reason draft pull_request body reactions performed_via_github_app state_reason repo type
268725471 MDU6SXNzdWUyNjg3MjU0NzE= 1662 Decoding time according to CF conventions raises error if a NaN is found 1956032 closed 0     4 2017-10-26T11:33:44Z 2017-11-21T14:38:41Z 2017-11-21T14:38:41Z CONTRIBUTOR      

Working with Argo data, I have difficulties decoding time-related variables: More specifically, it may happens that a variable being a date contains FillValue that are set to NaN at the opening of the netcdf file. That makes the decoding to raise an error.

Sure I can open the netcdf file with the decode_times = False option but it's not an issue of being able or not to decode the data, it seems to me to be about how to handle FillValue in a time axis.

I understand that with most of gridded datasets, the time axis/dimension/coordinate is full and does not contains missing values, that may be explaining why nobody have reported this before.

Here is a simple way to reproduce the error:

``` attrs = {'units': 'days since 1950-01-01 00:00:00 UTC'} # Classic Argo data Julian Day units

OK !

jd = [24658.46875, 24658.46366898, 24658.47256944] # Sample of Julian date from Argo data ds = xr.Dataset({'time': ('time', jd, attrs)}) print xr.decode_cf(ds)

<xarray.Dataset> Dimensions: (time: 3) Coordinates: * time (time) datetime64[ns] 2017-07-06T11:15:00 ... Data variables: empty But then:

Not OK with a NaN

jd = [24658.46875, 24658.46366898, 24658.47256944, np.NaN] # Another sample of Julian date from Argo data ds = xr.Dataset({'time': ('time', jd, attrs)}) print xr.decode_cf(ds)

ValueError: unable to decode time units 'days since 1950-01-01 00:00:00 UTC' with the default calendar. Try opening your dataset with decode_times=False. Full traceback: Traceback (most recent call last): File "/Users/gmaze/anaconda/envs/obidam/lib/python2.7/site-packages/xarray/conventions.py", line 389, in init result = decode_cf_datetime(example_value, units, calendar) File "/Users/gmaze/anaconda/envs/obidam/lib/python2.7/site-packages/xarray/conventions.py", line 157, in decode_cf_datetime dates = _decode_datetime_with_netcdf4(flat_num_dates, units, calendar) File "/Users/gmaze/anaconda/envs/obidam/lib/python2.7/site-packages/xarray/conventions.py", line 99, in _decode_datetime_with_netcdf4 dates = np.asarray(nc4.num2date(num_dates, units, calendar)) File "netCDF4/_netCDF4.pyx", line 5244, in netCDF4._netCDF4.num2date (netCDF4/_netCDF4.c:64839) ValueError: cannot convert float NaN to integer ```

I would expect the decoding to work like in the first case and to simply preserve NaNs where they are.

Any ideas or suggestions ? Thanks

 
{
    "url": "https://api.github.com/repos/pydata/xarray/issues/1662/reactions",
    "total_count": 0,
    "+1": 0,
    "-1": 0,
    "laugh": 0,
    "hooray": 0,
    "confused": 0,
    "heart": 0,
    "rocket": 0,
    "eyes": 0
}
   completed 13221727 issue

Links from other tables