User directories (~) not expanded by open_mfdataset().

[juseg]

What happened:

Hi. Perhaps a trivially minor issue, but I have noticed that open_dataset and to_netcdf expand user directories (~), while open_mfdataset does so on file lists, but not on patterns or single files.

What you expected to happen:

I would expect consistent behaviour across all file reading and writing methods.

Minimal Complete Verifiable Example:

>>> import xarray as xr
>>> xr.Dataset().to_netcdf('~/dataset.nc')  # works
>>> xr.open_dataset('~/dataset.nc')         # works
>>> xr.open_mfdataset(['~/dataset.nc'])     # works
>>> xr.open_mfdataset('~/dataset.nc')       # error
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3.9/site-packages/xarray/backends/api.py", line 891, in open_mfdataset
    raise OSError("no files to open")
OSError: no files to open

Anything else we need to know?:

This is done by private method _normalize_paths, which is called in open_dataset and to_netcdf but not open_mfdataset.

xarray/xarray/backends/api.py

Lines 170 to 177 in 1525fb0

	def _normalize_path(path):
	if isinstance(path, Path):
	path = str(path)
	if isinstance(path, str) and not is_remote_uri(path):
	path = os.path.abspath(os.path.expanduser(path))
	return path

Environment:

Output of xr.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.9.1 (default, Dec 13 2020, 11:55:53)
[GCC 10.2.0]
python-bits: 64
OS: Linux
OS-release: 5.4.80-2-MANJARO
machine: x86_64
processor:
byteorder: little
LC_ALL: None
LANG: en_GB.utf8
LOCALE: en_GB.UTF-8
libhdf5: 1.12.0
libnetcdf: 4.7.4

xarray: 0.16.1
pandas: 1.1.5
numpy: 1.19.4
scipy: 1.5.4
netCDF4: 1.5.4
pydap: None
h5netcdf: None
h5py: None
Nio: None
zarr: None
cftime: 1.2.1
nc_time_axis: None
PseudoNetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: None
dask: 2.30.0
distributed: None
matplotlib: 3.3.3
cartopy: None
seaborn: None
numbagg: None
pint: None
setuptools: 51.0.0
pip: 20.2.2
conda: None
pytest: 6.1.2
IPython: 7.19.0
sphinx: 3.4.0