html_url,issue_url,id,node_id,user,created_at,updated_at,author_association,body,reactions,performed_via_github_app,issue
https://github.com/pydata/xarray/issues/4765#issuecomment-755190063,https://api.github.com/repos/pydata/xarray/issues/4765,755190063,MDEyOklzc3VlQ29tbWVudDc1NTE5MDA2Mw==,34008825,2021-01-06T09:33:11Z,2021-01-06T13:45:01Z,NONE,"> I think that works as intended, although the error message could definitely be better.
> 
> `Series` is limited to 1D arrays, so we have to convert `DataArray` objects with more than one dimension using `stack`. If the object already has a `MultiIndex` this fails because we don't support `da.stack(xy=['x', 'y']).stack(xyz=['xy', 'z'])`.

It's a pity because having a partial MultiIndex DataArray helps a lot  if there are lots of combinations of `x` and `y` with NA values. My reason to transfort into a pandas object is that I need to `groupby` by 3 dimensions.

EDIT: 
I have found a strange workaround: `da.stack(xy=['x', 'y']).to_dask_dataframe().compute()`","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,778863478