Skip to content

Backport PR #35578 on branch 1.1.x (BUG: df.shift(n, axis=1) with multiple blocks) #35601

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.

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
Show file tree
Hide file tree
Changes from all commits
Commits
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
1 change: 1 addition & 0 deletions doc/source/whatsnew/v1.1.1.rst
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@ Fixed regressions

- Fixed regression where :func:`read_csv` would raise a ``ValueError`` when ``pandas.options.mode.use_inf_as_na`` was set to ``True`` (:issue:`35493`).
- Fixed regression in :class:`pandas.core.groupby.RollingGroupby` where column selection was ignored (:issue:`35486`)
- Fixed regression in :meth:`DataFrame.shift` with ``axis=1`` and heterogeneous dtypes (:issue:`35488`)

.. ---------------------------------------------------------------------------

Expand Down
25 changes: 23 additions & 2 deletions pandas/core/internals/managers.py
Original file line number Diff line number Diff line change
Expand Up @@ -551,6 +551,24 @@ def interpolate(self, **kwargs) -> "BlockManager":
return self.apply("interpolate", **kwargs)

def shift(self, periods: int, axis: int, fill_value) -> "BlockManager":
if axis == 0 and self.ndim == 2 and self.nblocks > 1:
# GH#35488 we need to watch out for multi-block cases
ncols = self.shape[0]
if periods > 0:
indexer = [-1] * periods + list(range(ncols - periods))
else:
nper = abs(periods)
indexer = list(range(nper, ncols)) + [-1] * nper
result = self.reindex_indexer(
self.items,
indexer,
axis=0,
fill_value=fill_value,
allow_dups=True,
consolidate=False,
)
return result

return self.apply("shift", periods=periods, axis=axis, fill_value=fill_value)

def fillna(self, value, limit, inplace: bool, downcast) -> "BlockManager":
Expand Down Expand Up @@ -1213,6 +1231,7 @@ def reindex_indexer(
fill_value=None,
allow_dups: bool = False,
copy: bool = True,
consolidate: bool = True,
) -> T:
"""
Parameters
Expand All @@ -1223,7 +1242,8 @@ def reindex_indexer(
fill_value : object, default None
allow_dups : bool, default False
copy : bool, default True

consolidate: bool, default True
Whether to consolidate inplace before reindexing.

pandas-indexer with -1's only.
"""
Expand All @@ -1236,7 +1256,8 @@ def reindex_indexer(
result.axes[axis] = new_axis
return result

self._consolidate_inplace()
if consolidate:
self._consolidate_inplace()

# some axes don't allow reindexing with dups
if not allow_dups:
Expand Down
27 changes: 27 additions & 0 deletions pandas/tests/frame/methods/test_shift.py
Original file line number Diff line number Diff line change
Expand Up @@ -145,6 +145,33 @@ def test_shift_duplicate_columns(self):
tm.assert_frame_equal(shifted[0], shifted[1])
tm.assert_frame_equal(shifted[0], shifted[2])

def test_shift_axis1_multiple_blocks(self):
# GH#35488
df1 = pd.DataFrame(np.random.randint(1000, size=(5, 3)))
df2 = pd.DataFrame(np.random.randint(1000, size=(5, 2)))
df3 = pd.concat([df1, df2], axis=1)
assert len(df3._mgr.blocks) == 2

result = df3.shift(2, axis=1)

expected = df3.take([-1, -1, 0, 1, 2], axis=1)
expected.iloc[:, :2] = np.nan
expected.columns = df3.columns

tm.assert_frame_equal(result, expected)

# Case with periods < 0
# rebuild df3 because `take` call above consolidated
df3 = pd.concat([df1, df2], axis=1)
assert len(df3._mgr.blocks) == 2
result = df3.shift(-2, axis=1)

expected = df3.take([2, 3, 4, -1, -1], axis=1)
expected.iloc[:, -2:] = np.nan
expected.columns = df3.columns

tm.assert_frame_equal(result, expected)

@pytest.mark.filterwarnings("ignore:tshift is deprecated:FutureWarning")
def test_tshift(self, datetime_frame):
# TODO: remove this test when tshift deprecation is enforced
Expand Down