xarray .count() gives -1 when it should be 1?

[ru111]

Data array da I'm working with looks like:

<xarray.DataArray 'T' (time: 365, latitude: 256, longitude: 512)>
[47841280 values with dtype=float32]
Coordinates:
  * time         (time) datetime64[ns] 1979-01-01T09:00:00 ... 1979-12-31T09:00:00
  * latitude     (latitude) float32 89.462944 88.766945 ... -89.462944
  * longitude    (longitude) float32 0.0 0.703125 ... 358.59375 359.29688

Now int(da.time.count()) gives 365 which is correct. However when I select a time coordinate da2 = da.isel(time=0) the resulting array has one time coordinate:

<xarray.DataArray 'T' (latitude: 256, longitude: 512)>
[131072 values with dtype=float32]
Coordinates:
    time         datetime64[ns] 1979-01-01T09:00:00
  * latitude     (latitude) float32 89.462944 88.766945 ... -89.462944
  * longitude    (longitude) float32 0.0 0.703125 ... 358.59375 359.29688

and int(da2.time.count()) gives me -1 instead of 1. I wasn't sure if this was a bug or I wasn't selecting the coordinate properly - if this is the case how should I call the time coordinate to count it correctly?

[max-sixty]

Thanks for the issue @ru111

That does seem weird. Do you have a minimally reproducible example?

[andrew-c-ross]

I can reproduce OP when the coordinate is a datetime64, but the problem does not occur with int64s (or float64s).

datetime coordinate:

In [1]: import xarray

In [2]: xarray.__version__
Out[2]: '0.11.2'

In [3]: import datetime as dt

In [4]: da = xarray.DataArray([0], coords=[('time', [dt.datetime(1979, 1, 1)])])

In [5]: da.isel(time=0).time.count()
Out[5]:
<xarray.DataArray 'time' ()>
array(-1)
Coordinates:
    time     datetime64[ns] 1979-01-01

int coordinate:

In [6]: da = xarray.DataArray([0], coords=[('time', [0])])

In [7]: da.isel(time=0).time.count()
Out[7]:
<xarray.DataArray 'time' ()>
array(1)
Coordinates:
    time     int64 0

[shoyer]

The problem seems to come up for scalar datetime64 arrays:

>>> xarray.Variable((), np.datetime64('2000-01-01')).count()
-1

[shoyer]

Apparently isnull(datetime64_scalar_array) returns a Python boolean False. This gets inverted by ~ to -1 instead of True in our implementation of count:

xarray/xarray/core/duck_array_ops.py

Line 188 in 4cd56a9

return np.sum(~isnull(data), axis=axis)

To fix this, we should either switch ~x to np.logical_not(x).

[dcherian]

Though for datetime64 we should use ~np.nat?