TYP: ExtensionArray unique and repeat #41260
Conversation
simonjayhawkins
added
Typing
|
Right, my thought is making those Literal would make it more obvious to
readers that it is for DataFrame
…On Sun, May 2, 2021 at 6:57 PM Irv Lustig ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In pandas/core/arrays/interval.py
<#41260 (comment)>:
> @@ -1518,7 +1518,11 @@ def delete(self: IntervalArrayT, loc) -> IntervalArrayT:
return self._shallow_copy(left=new_left, right=new_right)
@appender(_extension_array_shared_docs["repeat"] % _shared_docs_kwargs)
- def repeat(self: IntervalArrayT, repeats: int, axis=None) -> IntervalArrayT:
+ def repeat(
+ self: IntervalArrayT,
+ repeats: int | Sequence[int],
+ axis: int | None = None,
@jbrockmendel <https://github.com/jbrockmendel> No, we can't accept those
arguments in ExtensionArray.repeat(), because we call numpy.repeat
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#41260 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB5UM6DILM4G7I7URYA2SDTTLX7I7ANCNFSM4364DQ3A>
.
|
@jbrockmendel Except that there is no |
|
pd._typing.Axis is intended for use in Series/DataFrame methods, not in EA methods. But that is not obvious to all readers (as evidenced by the initial suggestion to use it here). So I was suggesting an edit that might make it clearer to the next reader. Out of scope for this PR. |
right my main point here is that we should have exactly 1 canonical way to type things and it should be obvious. If its not it needs lots of comments / expl and is not ideal at all, otherwise we are going to get inconsistencies across the board. if there is a way to enforce this would be great. in any event this PR is fine. |
Typing fixes for
ExtensionArray.unique()andExtensionArray.repeat()