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/3461#issuecomment-547620729,https://api.github.com/repos/pydata/xarray/issues/3461,547620729,MDEyOklzc3VlQ29tbWVudDU0NzYyMDcyOQ==,6213168,2019-10-29T20:43:34Z,2019-10-29T20:43:34Z,MEMBER,I think there's space here to make xarray more user friendly. A better error message would particularly be useful with new xarray users which are used to numpy and pandas - which is the most common new user profile TBH.,"{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,514191264
https://github.com/pydata/xarray/issues/3461#issuecomment-547614151,https://api.github.com/repos/pydata/xarray/issues/3461,547614151,MDEyOklzc3VlQ29tbWVudDU0NzYxNDE1MQ==,6213168,2019-10-29T20:27:14Z,2019-10-29T20:33:27Z,MEMBER,"@rpgoldman if you want to do an _elementwise sum_ of the data_vars of a Dataset, it's as simple as ``sum(ds.data_vars.values())``.
``Dataset.sum()`` is not elementwise - it's a reduction along one or more axes - exactly like in numpy.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,514191264
https://github.com/pydata/xarray/issues/3461#issuecomment-547613150,https://api.github.com/repos/pydata/xarray/issues/3461,547613150,MDEyOklzc3VlQ29tbWVudDU0NzYxMzE1MA==,6213168,2019-10-29T20:24:50Z,2019-10-29T20:24:50Z,MEMBER,"It's a case of improper error message.

I don't think anybody ever thought of passing  the ``axis`` parameter to any of the Dataset methods before. As a general rule, it would be a very unwise idea unless you're sure that *all* of your data_vars have the required dimension exactly in the same position.

Note the documentation for Dataset.sum vs. DataArray.sum:

http://xarray.pydata.org/en/stable/generated/xarray.Dataset.sum.html
http://xarray.pydata.org/en/stable/generated/xarray.DataArray.sum.html

You were able to pass the axis parameter because Dataset.sum accepts **kwargs, which are intended for backend-specific arguments (e.g. dask's ``split_every`` parameter).

The solution is to blacklist a bunch of numpy arguments that don't make sense (``axis`` is one, but ``out`` also comes to mind).","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,514191264