Skip to content

BUG: Groupby not keeping string dtype for empty objects #55619

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 4 commits into from
Oct 26, 2023
Merged

BUG: Groupby not keeping string dtype for empty objects #55619

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
0cb459c
BUG: Groupby not keeping string dtype for empty objects
phofl Oct 21, 2023
506c2b2
Fix
phofl Oct 21, 2023
bd1c07d
Merge branch 'main' into string_dtype_groupby_len_zero
phofl Oct 24, 2023
334dab9
Merge branch 'main' into string_dtype_groupby_len_zero
lithomas1 Oct 25, 2023
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/v2.1.2.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@ Fixed regressions

Bug fixes
~~~~~~~~~
- Fixed bug in :meth:`.DataFrameGroupBy.min()` and :meth:`.DataFrameGroupBy.max()` not preserving extension dtype for empty object (:issue:`55619`)
- Fixed bug in :meth:`Categorical.equals` if other has arrow backed string dtype (:issue:`55364`)
- Fixed bug in :meth:`DataFrame.__setitem__` not inferring string dtype for zero-dimensional array with ``infer_string=True`` (:issue:`55366`)
- Fixed bug in :meth:`DataFrame.idxmin` and :meth:`DataFrame.idxmax` raising for arrow dtypes (:issue:`55368`)
Expand Down
3 changes: 3 additions & 0 deletions pandas/core/arrays/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2352,6 +2352,9 @@ def _groupby_op(
# GH#43682
if isinstance(self.dtype, StringDtype):
# StringArray
if op.how not in ["any", "all"]:
# Fail early to avoid conversion to object
op._get_cython_function(op.kind, op.how, np.dtype(object), False)
npvalues = self.to_numpy(object, na_value=np.nan)
else:
raise NotImplementedError(
Expand Down
20 changes: 13 additions & 7 deletions pandas/core/groupby/ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,6 +33,7 @@
from pandas.errors import AbstractMethodError
from pandas.util._decorators import cache_readonly

from pandas.core.dtypes.base import ExtensionDtype
from pandas.core.dtypes.cast import (
maybe_cast_pointwise_result,
maybe_downcast_to_dtype,
Expand Down Expand Up @@ -837,10 +838,8 @@ def agg_series(
-------
np.ndarray or ExtensionArray
"""
# test_groupby_empty_with_category gets here with self.ngroups == 0
# and len(obj) > 0

if len(obj) > 0 and not isinstance(obj._values, np.ndarray):
if not isinstance(obj._values, np.ndarray):
# we can preserve a little bit more aggressively with EA dtype
# because maybe_cast_pointwise_result will do a try/except
# with _from_sequence. NB we are assuming here that _from_sequence
Expand All @@ -849,11 +848,18 @@ def agg_series(

result = self._aggregate_series_pure_python(obj, func)

npvalues = lib.maybe_convert_objects(result, try_float=False)
if preserve_dtype:
out = maybe_cast_pointwise_result(npvalues, obj.dtype, numeric_only=True)
if len(obj) == 0 and len(result) == 0 and isinstance(obj.dtype, ExtensionDtype):
Copy link
Member

Choose a reason for hiding this comment

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

What happens on an empty list of categoricals with observed=False? I think this is the only case where len(obj) == 0 but len(result) > 0.

Copy link
Member Author

Choose a reason for hiding this comment

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

I didn't track his down specifically, but the test that was mentioned in the comment is not passing by here anymore

Copy link
Member

Choose a reason for hiding this comment

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

Running on this PR as-is:

func = "min"
dtype = "string[pyarrow_numpy]"
df = DataFrame({"a": ["a"], "b": "a", "c": "a"}, dtype=dtype)
df["a"] = pd.Categorical([1], categories=[1, 2, 3])
df = df.iloc[:0]
result = getattr(df.groupby("a", observed=False), func)()

print(result.dtypes)
# b    float64
# c    float64
# dtype: object

If you remove the condition len(result) == 0 in the line highlighted above, you get string[pyarrow_numpy] for the dtypes instead. However, the values in the DataFrame are still the float NaN. I didn't realize this was possible?

Copy link
Member Author

Choose a reason for hiding this comment

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

The new string dtype uses nan as missing value representation to keep numpy semantics, so it will work if you feed it an array with all NaN

Copy link
Member

Choose a reason for hiding this comment

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

Ah - thanks

Copy link
Member

Choose a reason for hiding this comment

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

I think this still needs fixed (but okay not for 2.1.2)

cls = obj.dtype.construct_array_type()
out = cls._from_sequence(result)

else:
out = npvalues
npvalues = lib.maybe_convert_objects(result, try_float=False)
if preserve_dtype:
out = maybe_cast_pointwise_result(
npvalues, obj.dtype, numeric_only=True
)
else:
out = npvalues
return out

@final
Expand Down
13 changes: 13 additions & 0 deletions pandas/tests/groupby/test_reductions.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -575,6 +575,19 @@ def test_groupby_min_max_categorical(func):
tm.assert_frame_equal(result, expected)


@pytest.mark.parametrize("func", ["min", "max"])
def test_min_empty_string_dtype(func):
# GH#55619
pytest.importorskip("pyarrow")
dtype = "string[pyarrow_numpy]"
df = DataFrame({"a": ["a"], "b": "a", "c": "a"}, dtype=dtype).iloc[:0]
result = getattr(df.groupby("a"), func)()
expected = DataFrame(
columns=["b", "c"], dtype=dtype, index=pd.Index([], dtype=dtype, name="a")
)
tm.assert_frame_equal(result, expected)


def test_max_nan_bug():
raw = """,Date,app,File
-04-23,2013-04-23 00:00:00,,log080001.log
Expand Down