Copied DataArray loses its set CRS

[apiwat-chantawibul]

Code Sample

import xarray as xr
import rioxarray as rxr


da = xr.DataArray(
    [[1,2],[3,4]],
    dims = ('y','x'),
    coords = dict(
        y = [0,1],
        x = [0,1],
    ),
)
da.rio.set_crs('EPSG:4326')
print('CRS before copy:', da.rio.crs)
da_ = da.copy()
print('CRS after copy:', da_.rio.crs)

Problem description

Currently it outputs:

CRS before copy: EPSG:4326
CRS after copy: None

Meaning CRS information is lost by .copy()

Expected Output

CRS before copy: EPSG:4326
CRS after copy: EPSG:4326

Environment Information

Tested on

rioxarray 0.10.3
python 3.10.4

[apiwat-chantawibul]

Viewing in another way, maybe the problem has to do with how .set_crs() sets an internal variable instead of something like .attrs which would have been copied automatically.

[apiwat-chantawibul]

Okay, I just found rio.write_crs() that explains a lot. So rio.set_crs() has to be followed by rio.write_crs() to make the CRS change portable. And then there is also rio.write_coordinate_system() for some reason. All this seems like a great trap for new users to me. Right now I'm just keeping this issue open in case anyone want to discuss why it has to be this way and how to possibly improve it.

[snowman2]

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

[snowman2]

Related #356