REGR: groupby with Decimal and NA values (pandas-dev#56522)

Author: rhshadrach

doc/source/whatsnew/v2.2.0.rst

@@ -659,11 +659,11 @@ Groupby/resample/rolling
 - Bug in :meth:`.DataFrameGroupBy.value_counts` and :meth:`.SeriesGroupBy.value_count` would sort by proportions rather than frequencies when ``sort=True`` and ``normalize=True`` (:issue:`55951`)
 - Bug in :meth:`DataFrame.asfreq` and :meth:`Series.asfreq` with a :class:`DatetimeIndex` with non-nanosecond resolution incorrectly converting to nanosecond resolution (:issue:`55958`)
 - Bug in :meth:`DataFrame.ewm` when passed ``times`` with non-nanosecond ``datetime64`` or :class:`DatetimeTZDtype` dtype (:issue:`56262`)
+- Bug in :meth:`DataFrame.groupby` and :meth:`Series.groupby` where grouping by a combination of ``Decimal`` and NA values would fail when ``sort=True`` (:issue:`54847`)
 - Bug in :meth:`DataFrame.resample` not respecting ``closed`` and ``label`` arguments for :class:`~pandas.tseries.offsets.BusinessDay` (:issue:`55282`)
 - Bug in :meth:`DataFrame.resample` when resampling on a :class:`ArrowDtype` of ``pyarrow.timestamp`` or ``pyarrow.duration`` type (:issue:`55989`)
 - Bug in :meth:`DataFrame.resample` where bin edges were not correct for :class:`~pandas.tseries.offsets.BusinessDay` (:issue:`55281`)
 - Bug in :meth:`DataFrame.resample` where bin edges were not correct for :class:`~pandas.tseries.offsets.MonthBegin` (:issue:`55271`)
--
 
 Reshaping
 ^^^^^^^^^

pandas/core/algorithms.py

@@ -4,6 +4,7 @@
 """
 from __future__ import annotations
 
+import decimal
 import operator
 from textwrap import dedent
 from typing import (
@@ -1514,7 +1515,7 @@ def safe_sort(
         try:
             sorter = values.argsort()
             ordered = values.take(sorter)
-        except TypeError:
+        except (TypeError, decimal.InvalidOperation):
             # Previous sorters failed or were not applicable, try `_sort_mixed`
             # which would work, but which fails for special case of 1d arrays
             # with tuples.

pandas/tests/groupby/test_groupby.py

@@ -1,4 +1,5 @@
 from datetime import datetime
+import decimal
 from decimal import Decimal
 import re
 
@@ -3313,3 +3314,23 @@ def test_depr_grouper_attrs(attr):
     msg = f"{attr} is deprecated"
     with tm.assert_produces_warning(FutureWarning, match=msg):
         getattr(gb.grouper, attr)
+
+
+@pytest.mark.parametrize("test_series", [True, False])
+def test_decimal_na_sort(test_series):
+    # GH#54847
+    # We catch both TypeError and decimal.InvalidOperation exceptions in safe_sort.
+    # If this next assert raises, we can just catch TypeError
+    assert not isinstance(decimal.InvalidOperation, TypeError)
+    df = DataFrame(
+        {
+            "key": [Decimal(1), Decimal(1), None, None],
+            "value": [Decimal(2), Decimal(3), Decimal(4), Decimal(5)],
+        }
+    )
+    gb = df.groupby("key", dropna=False)
+    if test_series:
+        gb = gb["value"]
+    result = gb.grouper.result_index
+    expected = Index([Decimal(1), None], name="key")
+    tm.assert_index_equal(result, expected)