BUG: Fix BaseWindowGroupby.aggregate where as_index is ignored #54973
Conversation
pandas/core/window/rolling.py
Outdated
| @@ -875,6 +875,14 @@ def _gotitem(self, key, ndim, subset=None): | |||
| subset = self.obj.set_index(self._on) | |||
| return super()._gotitem(key, ndim, subset=subset) | |||
|
|
|||
| def aggregate(self, func, *args, **kwargs): | |||
| result = super().aggregate(func, *args, **kwargs) | |||
| if not self._as_index: | |||
There was a problem hiding this comment.
I think this should live in BaseWindowGroupby so added this aggregate function. After adding, mypy complained about incompatibility of base classes for classes inheriting BaseWindowGroupby. Adding aggregate function to those solves the issue happy to extend their docs if required or implement another workaround if preferred over this
pandas/core/window/ewm.py
Outdated
| @@ -919,6 +919,12 @@ def _get_window_indexer(self) -> GroupbyIndexer: | |||
| ) | |||
| return window_indexer | |||
|
|
|||
| @doc(_shared_docs["aggregate"]) | |||
There was a problem hiding this comment.
Shouldn't this and below be inherited from the base class?
There was a problem hiding this comment.
I think so. This was an ugly workaround to a mypy error I was getting but using type: ignore[misc] instead now
doc/source/whatsnew/v2.2.0.rst
Outdated
| @@ -168,6 +168,7 @@ Performance improvements | |||
| Bug fixes | |||
| ~~~~~~~~~ | |||
| - Bug in :class:`AbstractHolidayCalendar` where timezone data was not propagated when computing holiday observances (:issue:`54580`) | |||
| - Bug in :meth:`BaseWindowGroupby.aggregate` where ``_as_index`` attribute is ignored (:issue:`31007`) | |||
There was a problem hiding this comment.
Could you write this in terms of public APIs?
mroeschke
added
Groupby
Apply
|
Also realized that this is failing with a list-like parameter as well so extended testing |
doc/source/whatsnew/v2.2.0.rstfile if fixing a bug or adding a new feature.