Skip to content

BUG: groupby.transform retains timezone information #25264

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 12 commits into from
Feb 16, 2019
Merged

BUG: groupby.transform retains timezone information #25264

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
4c7e66d
BUG: groupby.transform retains timezone information
Feb 11, 2019
3c5fd19
Add whatsnew
Feb 11, 2019
032e201
Move whatsnew to 0.24.2
Feb 11, 2019
a7678a3
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 11, 2019
ab34a5c
Add more tests
Feb 11, 2019
dfe6124
Add first last
Feb 12, 2019
8f808ce
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 12, 2019
dee41d3
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 13, 2019
ea016f0
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 14, 2019
8b6a8e6
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 14, 2019
6f06d83
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 15, 2019
288aa07
Merge remote-tracking branch 'upstream/master' into timezone_transform
Feb 16, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion doc/source/whatsnew/v0.24.2.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -76,7 +76,7 @@ Bug Fixes

**Reshaping**

-
- Bug in :meth:`pandas.core.groupby.GroupBy.transform` where applying a function to a timezone aware column would return a timezone naive result (:issue:`24198`)
-
-

Expand Down
2 changes: 1 addition & 1 deletion pandas/core/groupby/generic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -964,7 +964,7 @@ def _transform_fast(self, func, func_nm):

ids, _, ngroup = self.grouper.group_info
cast = self._transform_should_cast(func_nm)
out = algorithms.take_1d(func().values, ids)
out = algorithms.take_1d(func()._values, ids)
if cast:
out = self._try_cast(out, self.obj)
return Series(out, index=self.obj.index, name=self.obj.name)
Expand Down
10 changes: 10 additions & 0 deletions pandas/tests/groupby/test_transform.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -834,3 +834,13 @@ def demean_rename(x):
tm.assert_frame_equal(result, expected)
result_single = df.groupby('group').value.transform(demean_rename)
tm.assert_series_equal(result_single, expected['value'])


def test_groupby_transform_timezone_column():
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure if it's too much here but might be worth creating a fixture for transformation functions and parametrizing based off of that here. That way we could ensure that the tz info doesn't get lost across all instead of just max.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fair point. For this test though, we need reducing functions that are applicable for timestamps though. All that make sense to me is min and max unless I am forgetting any other applicable functions.

If the list is short, may be worth just parameterizing over directly instead?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yea agreed with your approach. I think it's worth while to start fixturizing these tests to ensure full coverage on particular subsets of algorithms, just don't know what those subsets should all be just yet. I think the way you've done it makes that refactor more apparent in the future if we ever get there

# GH 24198
ts = pd.to_datetime('now', utc=True).tz_convert('Asia/Singapore')
result = pd.DataFrame({'end_time': [ts], 'id': [1]})
result['max_end_time'] = result.groupby('id').end_time.transform(max)
expected = pd.DataFrame([[ts, 1, ts]], columns=['end_time', 'id',
'max_end_time'])
tm.assert_frame_equal(result, expected)