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/862#issuecomment-224949305,https://api.github.com/repos/pydata/xarray/issues/862,224949305,MDEyOklzc3VlQ29tbWVudDIyNDk0OTMwNQ==,6079398,2016-06-09T16:26:36Z,2016-06-09T16:26:36Z,NONE,"> This seems a little too magical to me. How would we know if the dataset dimension was added intentionally or not? Yeah, that's a fair point. I'll put together something that uses an optional list of dimensions to concatenate over. Thanks! ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,157545837 https://github.com/pydata/xarray/issues/862#issuecomment-224787831,https://api.github.com/repos/pydata/xarray/issues/862,224787831,MDEyOklzc3VlQ29tbWVudDIyNDc4NzgzMQ==,6079398,2016-06-09T02:51:11Z,2016-06-09T02:51:52Z,NONE,"Hey @shoyer, Sorry for the delayed response. Passing a list of dimensions over which to concatenate over seems like it would be the easiest workaround with the fewest questions asked. As you mentioned, every dimension gets a variable by the time it is a dataset, so another option (that I'll admit I haven't thought all the way through and may not even work) would be to first check if `decode_cf` is working on a `Dataset` or `AbstractDataStore` (e: which it already does anyway), and then decide whether to concatenate over a dimension or not. I could see the latter idea not working out so well, but I'd be curious about your thoughts. Either way, I can put something together this week and open up a PR. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,157545837