Resampling daily input data to half-yearly data generates an excessive time coordinate

[grassland-curing-cfa]

Code Sample, a copy-pastable example if possible

A "Minimal, Complete and Verifiable Example" will make it much easier for maintainers to help you:
http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports

# Your code here
import numpy as np
import pandas as pd
import xarray as xr

time = pd.date_range('1972-01-01', freq='D', periods=366)    # a leap year
ds = xr.Dataset({'foo': ('time', np.arange(366)), 'time': time})

# Resample ds to every 6 months
res = ds['foo'].resample(time='6MS', closed='left').sum('time')
print(res)

Problem description

<xarray.DataArray 'foo' (time: 3)>
array([16471., 50324., nan])
Coordinates:

• time (time) datetime64[ns] 1972-01-01 1972-07-01 1973-01-01

An excessive time coordinate of 1973-01-01 was generated despite the variable values being nan.

Expected Output

<xarray.DataArray 'foo' (time: 2)>
array([16471., 50324.])
Coordinates:

• time (time) datetime64[ns] 1972-01-01 1972-07-01

Output of xr.show_versions()

# Paste the output here xr.show_versions() here

python: 3.6.5 (v3.6.5:f59c0932b4, Mar 28 2018, 17:00:18) [MSC v.1900 64 bit (AMD64)]
python-bits: 64
OS: Windows
OS-release: 2012ServerR2
machine: AMD64
processor: Intel64 Family 6 Model 47 Stepping 2, GenuineIntel

xarray: 0.11.2
pandas: 0.23.4
numpy: 1.16.0
scipy: 1.2.0
netCDF4: 1.4.2

[spencerkclark]

We've noticed this behavior before too: #2593 (comment). It turns out it was fixed in the most recent version of pandas, so I recommend upgrading:

In [1]: import pandas as pd

In [2]: pd.__version__
Out[2]: '0.24.0'

In [3]: import xarray as xr

In [4]: time = pd.date_range('1972-01-01', freq='D', periods=366)

In [5]: ds = xr.Dataset({'foo': ('time', range(366)), 'time': time})

In [6]: res = ds['foo'].resample(time='6MS', closed='left').sum('time')

In [7]: res
Out[7]:
<xarray.DataArray 'foo' (time: 2)>
array([16471, 50324])
Coordinates:
  * time     (time) datetime64[ns] 1972-01-01 1972-07-01

[grassland-curing-cfa]

Thanks @spencerkclark ! This fix worked for me.