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/4386#issuecomment-808831290,https://api.github.com/repos/pydata/xarray/issues/4386,808831290,MDEyOklzc3VlQ29tbWVudDgwODgzMTI5MA==,13301940,2021-03-28T01:58:27Z,2021-03-28T02:00:31Z,MEMBER,"I think the culprit here is 

https://github.com/pydata/xarray/blob/2bbac154ddacd6d6ec989b040130cd9afb8dc0a1/xarray/conventions.py#L136-L147

As I understand it, 

- This approach was adopted in https://github.com/pydata/xarray/pull/849 to circumvent https://github.com/Unidata/netcdf-c/issues/267. 

- The round-tripping fails due to `xr.open_zarr('/tmp/ds1.zarr')` resulting in a dataset with 'x' variable with an encoding that has a `'dtype': dtype('int8')` entry. 

-  since zarr doesn't have a problem with serializing booleans, **should we special-case** zarr and other backends with native support for booleans instead of casting to `int8`? 
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,688501399