CRS set on Dataset does not persist for DataArrays obtained by selecting data variable

[vidurmithal]

Code Sample

da1 = xr.DataArray(np.arange(10*10).reshape(10, 10), [("x", np.linspace(70,80,10)), ("y", np.linspace(20,30,10))], name='var_1')
da2 = xr.DataArray(np.arange(10*10).reshape(10, 10), [("x", np.linspace(70,80,10)), ("y", np.linspace(20,30,10))], name='var_2')

ds = xr.merge([da1,da2])

ds = ds.rio.set_crs('epsg:4326')

print('Dataset CRS: {}'.format(ds.rio.crs))
print('DataArray 1 CRS: {}'.format(ds['var_1'].rio.crs))
print('DataArray 2 CRS: {}'.format(ds['var_2'].rio.crs))

Problem description

The CRS set on a Dataset using .rio.set_crs() does not persist to a DataArray obtained by selecting individual variables from the Dataset. Printing the CRS of the Dataset gives the previously set CRS, but printing the CRS of the DataArray gives None.

Expected Output

The CRS set on a Dataset should also apply to DataArrays of individual variables. Printing DataArray.rio.crs should give the set CRS, EPSG:4326 in this case.

Environment Information

rioxarray (0.1.1) deps:
  rasterio: 1.1.0
    xarray: 0.16.2
      GDAL: 3.0.2

Other python deps:
     scipy: 1.4.1
    pyproj: 2.6.0

System:
   python: 3.7.7 (default, Mar 23 2020, 22:36:06)  [GCC 7.3.0]
   machine: Linux-5.4.89+-x86_64-with-debian-bullseye-sid

Installation method

PyPi

[snowman2]

We recommend using rio.write_crs as it persists better.

[snowman2]

Related #252

[snowman2]

For those finding this issue, I recommend reading the Getting Started: Introductory Information references for rioxarray. The Coordinare Reference System Management section is relevant for this issue.