BUG: drop_duplicates drops name(s). #10116
Conversation
| @@ -2580,7 +2580,10 @@ def drop(self, labels, errors='raise'): | |||
| @Appender(_shared_docs['drop_duplicates'] % _index_doc_kwargs) | |||
| def drop_duplicates(self, take_last=False): | |||
| result = super(Index, self).drop_duplicates(take_last=take_last) | |||
| return self._constructor(result) | |||
| if self.name is not None: | |||
There was a problem hiding this comment.
This check is for dealing with the API differences between Index and MultiIndex, yes? I think it would be cleaner to override the method on MultiIndex.
There was a problem hiding this comment.
Yes. OK, I can do that.
Alternatively, it would be great if the constructors of all Index subclasses (other than MultiIndex) accepted a names list of length one and used it to set name. That way one could always just set names without worrying about it.
There was a problem hiding this comment.
I would even be happy if you just explicitly check isinstance(self, MultiIndex).
I agree that your alternative solution would be preferable. That's part of a much bigger issue: #3268
jreback
added
Bug
Indexing
| if self.name is not None: | ||
| return self._constructor(result, name=self.name) | ||
| else: | ||
| return self._constructor(result, names=self.names) |
There was a problem hiding this comment.
you can dispense with this with a
return self._shallow_copy(result), should handle all meta-data
e.g. this will propogate things like freq/tz, etc as well.
There was a problem hiding this comment.
though looking at this now, I think MultiIndex._shallow_copy is wrong
let me see if I can fix this
|
@seth-p can you rebase and update? |
|
Afraid I do not have time to work on this. Could someone else handle it? |
|
@seth-p OK, will do. |
Closes #10115.