BUG: Fix groupby nth with axis=1 #43926
Conversation
|
Hello @johnzangwill! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found: There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻 Comment last updated at 2021-10-13 07:59:55 UTC |
| out = out.reindex(result_index) | ||
| out = self._reindex_output(out) | ||
| else: | ||
| out.columns = result_index[ids[mask]] |
There was a problem hiding this comment.
can we re-use any of the _wrap_foo_result methods?
pandas/core/groupby/groupby.py
Outdated
| @@ -2545,18 +2545,25 @@ def nth( | |||
| # Drop NA values in grouping | |||
| mask = mask & (ids != -1) | |||
|
|
|||
| out = self._selected_obj[mask] | |||
| if self.axis == 0: | |||
There was a problem hiding this comment.
actually let's make a method
self._selected_obj_for_axis(mask: np.ndarray[bool] | None)
as i think we use this in several places.
There was a problem hiding this comment.
actually let's make a method
self._selected_obj_for_axis(mask: np.ndarray[bool] | None)as i think we use this in several places.
@jreback, I had already factored these out in my other branch but I have done it here too, and made them both _mask_selected_obj(mask)
…zangwill/pandas into fix_groupby_nth_column_axis
Done |
doc/source/whatsnew/v1.4.0.rst
Outdated
| @@ -503,6 +503,7 @@ Groupby/resample/rolling | |||
| - Bug in :meth:`GroupBy.apply` with time-based :class:`Grouper` objects incorrectly raising ``ValueError`` in corner cases where the grouping vector contains a ``NaT`` (:issue:`43500`, :issue:`43515`) | |||
| - Bug in :meth:`GroupBy.mean` failing with ``complex`` dtype (:issue:`43701`) | |||
| - Fixed bug in :meth:`Series.rolling` and :meth:`DataFrame.rolling` not calculating window bounds correctly for the first row when ``center=True`` and index is decreasing (:issue:`43927`) | |||
| - Bug in :meth:`GroupBy.nth` failing on column groupings (``axis=1``) | |||
|
@jreback all green apart from |
Still all green apart from another quite frequent ci testing failure: Which I am sure is nothing to do with my code... |
doc/source/whatsnew/v1.4.0.rst
Outdated
| @@ -505,6 +505,7 @@ Groupby/resample/rolling | |||
| - Bug in :meth:`GroupBy.apply` with time-based :class:`Grouper` objects incorrectly raising ``ValueError`` in corner cases where the grouping vector contains a ``NaT`` (:issue:`43500`, :issue:`43515`) | |||
| - Bug in :meth:`GroupBy.mean` failing with ``complex`` dtype (:issue:`43701`) | |||
| - Fixed bug in :meth:`Series.rolling` and :meth:`DataFrame.rolling` not calculating window bounds correctly for the first row when ``center=True`` and index is decreasing (:issue:`43927`) | |||
| - Bug in :meth:`GroupBy.nth` failing on ``axis=1`` | |||
There was a problem hiding this comment.
can you add the PR number here (as the issue number)
There was a problem hiding this comment.
can you add the PR number here (as the issue number)
Done
| return self._mask_selected_obj(mask) | ||
|
|
||
| @final | ||
| def _mask_selected_obj(self, mask: np.ndarray) -> NDFrameT: | ||
| if self.axis == 0: |
There was a problem hiding this comment.
can you add a doc-string here
There was a problem hiding this comment.
can you add a doc-string here
Done
|
thanks @johnzangwill very nice |
| return self._mask_selected_obj(mask) | ||
|
|
||
| @final | ||
| def _mask_selected_obj(self, mask: np.ndarray) -> NDFrameT: |
Fixes a bug that no one has reported and probably no one cares about. But overlaps slightly with #42947, so I thought that I would make it a separate PR.
Current behaviour:
df.groupby(df.iloc[1], axis=1).nth(0)crashes ungracefully.