Function warning when sequence is all np.nan

What is your issue?

When processing the data, even with the bottleneck package installed, there is still a Warning when the sequence is empty. is it because bottleneck can't handle the .quantile function and still ends up calling the numpy function?

bottleneck: speeds up NaN-skipping and rolling window aggregations by a large factor

import xarray as xr
import numpy as np
array = np.random.randn(3,5,5)
array[:,2,2]=np.nan
data = xr.DataArray(array,
                    coords={'time': np.arange(3),
                            'lon': np.arange(5),
                            'lat': np.arange(5)}
                            )
data.quantile(0.95,dim='time', method='median_unbiased')

/home/yu/miniconda3/envs/learn/lib/python3.11/site-packages/numpy/lib/nanfunctions.py:1563: RuntimeWarning: All-NaN slice encountered
  return function_base._ureduce(a,

Start by reproducing the provided xarray DataArray example with the bottleneck package installed and confirm the RuntimeWarning from NumPy's nanfunctions.py. Trace the DataArray.quantile path to determine how all-NaN slices are handled; done means the intended quantile behavior is established and the warning behavior is covered by a regression test.