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/1628#issuecomment-336488363,https://api.github.com/repos/pydata/xarray/issues/1628,336488363,MDEyOklzc3VlQ29tbWVudDMzNjQ4ODM2Mw==,1217238,2017-10-13T15:37:21Z,2017-10-13T15:37:21Z,MEMBER,"> This type of data access is convenient but sometimes conflicts to the original attributes, e.g. T cannot be accessed like this.

The Dataset `T` property for transpose should be deprecated -- but that's a separate matter:
https://github.com/pydata/xarray/issues/1232

> Can we gradually deprecate this after implementing _ipython_key_completions_()?

I am hesitant to deprecate this since it is so widely used. But you're right that this serves the main use cases.

> An attribute foo can be accessed only by attribute-style, but not `__getitem__` style, `da[foo]`.

That's correct. Attribute-style access was always intended as fall-back for convenience purposes, but `__getitem__` (for programmatic access) is intentionally stricter.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,264764526