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/1326#issuecomment-289559455,https://api.github.com/repos/pydata/xarray/issues/1326,289559455,MDEyOklzc3VlQ29tbWVudDI4OTU1OTQ1NQ==,1217238,2017-03-27T19:28:36Z,2017-03-27T19:28:36Z,MEMBER,"One thing to make sure is if `dim` is already a scalar coordinate, it should be promoted to a 1D coordinate consisting of a single value.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,216799807
https://github.com/pydata/xarray/issues/1326#issuecomment-289520057,https://api.github.com/repos/pydata/xarray/issues/1326,289520057,MDEyOklzc3VlQ29tbWVudDI4OTUyMDA1Nw==,1217238,2017-03-27T17:12:45Z,2017-03-27T17:12:45Z,MEMBER,"Looks good to me. I would just make the default `axis=None` and indicate that that means ""insert all new dimension in order at the start of the result arrays"".","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,216799807
https://github.com/pydata/xarray/issues/1326#issuecomment-289513566,https://api.github.com/repos/pydata/xarray/issues/1326,289513566,MDEyOklzc3VlQ29tbWVudDI4OTUxMzU2Ng==,1217238,2017-03-27T16:50:02Z,2017-03-27T16:50:02Z,MEMBER,"> Do you think it is confusing if this method automatically detects the positional argument, i.e. if string is passed then it should be the axis name and an integer should be axis position.

Yes, I think this would be too confusing. Function arguments should do one thing.

What would `da.expand_dims(axis=0)` even do? Insert a new dimension with an auto-generated name like `dim_0`? That seems not very useful to me.

> I guess that the reason that numpy doesn't support multiple axes insertion is it brings additional ambugousity to the new axes positions;

I think we could clarify this is by saying that `axis` indicates axis positions on the *result* array, not the input array, rather than corresponding to repeated application of `expand_dims`. So `da.expand_dims(dim=['x' ,'y'], axis=[2, 1])` indicates that axis 2 on the result should be `x` and axis 1 should be `y`.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,216799807
https://github.com/pydata/xarray/issues/1326#issuecomment-289497437,https://api.github.com/repos/pydata/xarray/issues/1326,289497437,MDEyOklzc3VlQ29tbWVudDI4OTQ5NzQzNw==,1217238,2017-03-27T15:54:18Z,2017-03-27T15:54:18Z,MEMBER,"@fujiisoup I like your proposed API, but I would switch the order of the method arguments. That way, users who don't care about axis position can use a positional argument.

We might also consider supporting inserting multiple axes at once, e.g., by passing a list/tuple to `dim`/`axis`. This would be easy to add later, though.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,216799807
https://github.com/pydata/xarray/issues/1326#issuecomment-289084136,https://api.github.com/repos/pydata/xarray/issues/1326,289084136,MDEyOklzc3VlQ29tbWVudDI4OTA4NDEzNg==,1217238,2017-03-24T17:10:19Z,2017-03-24T17:10:19Z,MEMBER,Can you propose what the API should look like?,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,216799807