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/3035#issuecomment-610116325,https://api.github.com/repos/pydata/xarray/issues/3035,610116325,MDEyOklzc3VlQ29tbWVudDYxMDExNjMyNQ==,35968931,2020-04-07T01:08:24Z,2020-04-07T06:42:15Z,MEMBER,"It should be possible to implement a new backend in xarray now - we've already done it for [multiple file formats](http://xarray.pydata.org/en/stable/installing.html#for-netcdf-and-io).

However, I suspect that if we completed the [project](https://github.com/pydata/xarray/projects/3) to provide a [Flexible API for Storage Backends](https://github.com/pydata/xarray/issues/1970), then the path to this would become simpler. At the moment each new storage format has to be special-cased within xarray code itself (e.g. zarr), but we want to get to a position where that I/O responsibility is delegated out to other libraries, probably via an [entrypoint](https://github.com/pydata/xarray/pull/3166). Then it would be much easier (and quicker) to develop the necessary interfaces for the formats you mention.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,458236359