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/pull/3649#issuecomment-568142720,https://api.github.com/repos/pydata/xarray/issues/3649,568142720,MDEyOklzc3VlQ29tbWVudDU2ODE0MjcyMA==,4801430,2019-12-21T02:03:21Z,2019-12-21T02:03:21Z,CONTRIBUTOR,Done,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,540601428
https://github.com/pydata/xarray/pull/3649#issuecomment-568109804,https://api.github.com/repos/pydata/xarray/issues/3649,568109804,MDEyOklzc3VlQ29tbWVudDU2ODEwOTgwNA==,4801430,2019-12-20T22:23:56Z,2019-12-20T22:23:56Z,CONTRIBUTOR,Will do!,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,540601428
https://github.com/pydata/xarray/pull/3649#issuecomment-568109114,https://api.github.com/repos/pydata/xarray/issues/3649,568109114,MDEyOklzc3VlQ29tbWVudDU2ODEwOTExNA==,35968931,2019-12-20T22:21:17Z,2019-12-20T22:21:17Z,MEMBER,">Yeah this ""automatic alignment"" behaviour is everywhere so I think it makes sense for it to be the default.

Okay, in that case shall we add an option to opt-in to the stricter check by passing `fill_value=None`? With a quick note about it in the docstring for that argument?","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,540601428
https://github.com/pydata/xarray/pull/3649#issuecomment-568106485,https://api.github.com/repos/pydata/xarray/issues/3649,568106485,MDEyOklzc3VlQ29tbWVudDU2ODEwNjQ4NQ==,2448579,2019-12-20T22:11:02Z,2019-12-20T22:11:02Z,MEMBER,"> Imagine if someone passes a bunch of datasets that they think form a hypercube but don't, then with the PR they would get returned a dataset with a panel of NaNs that they would not be explicitly told are there... On the other hand xarray's other combining functions (e.g. merge and concat) will just insert NaNs.

Yeah this ""automatic alignment"" behaviour is everywhere so I think it makes sense for it to be the default.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,540601428
https://github.com/pydata/xarray/pull/3649#issuecomment-568067947,https://api.github.com/repos/pydata/xarray/issues/3649,568067947,MDEyOklzc3VlQ29tbWVudDU2ODA2Nzk0Nw==,4801430,2019-12-20T19:56:19Z,2019-12-20T19:56:56Z,CONTRIBUTOR,"yeah - that makes sense! So if `fill_value is None` then we can have the additional check to make sure it's a complete hypercube. The next question, like you mention, is whether we change the default `fill_value` to `None`. It seems like that would make the most sense to me. That way, the default is to throw an error with incomplete hypercubes (the previous behavior), and the incompleteness is only allowed if you provide a value to fill with. But I can see your point about having the default be `dtypes.NA` for consistency with `merge` and `concat`. Thoughts?","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,540601428