BUG: groupby-nunique modifies null values

doc/source/whatsnew/v1.0.2.rst

@@ -19,7 +19,7 @@ Fixed regressions
 - Fixed regression in :meth:`Series.align` when ``other`` is a DataFrame and ``method`` is not None (:issue:`31785`)
 - Fixed regression in :meth:`pandas.core.groupby.RollingGroupby.apply` where the ``raw`` parameter was ignored (:issue:`31754`)
 - Fixed regression in :meth:`rolling(..).corr() <pandas.core.window.Rolling.corr>` when using a time offset (:issue:`31789`)
--
+- Fixed regression in :meth:`DataFrameGroupBy.nunique` which was modifying the original values if ``NaN`` values were present (:issue:`31950`)
 
 .. ---------------------------------------------------------------------------
 

pandas/core/groupby/generic.py

@@ -591,20 +591,24 @@ def nunique(self, dropna: bool = True) -> Series:
 
         val = self.obj._internal_get_values()
 
-        # GH 27951
-        # temporary fix while we wait for NumPy bug 12629 to be fixed
-        val[isna(val)] = np.datetime64("NaT")
-
-        try:
-            sorter = np.lexsort((val, ids))
-        except TypeError:  # catches object dtypes
-            msg = f"val.dtype must be object, got {val.dtype}"
-            assert val.dtype == object, msg
+        def _object_sorter(val, ids):
             val, _ = algorithms.factorize(val, sort=False)
             sorter = np.lexsort((val, ids))
             _isna = lambda a: a == -1
+            return val, sorter, _isna
+
+        if isna(val).any() and val.dtype == object:
+            # Deal with pandas.NaT
+            val, sorter, _isna = _object_sorter(val, ids)
         else:
-            _isna = isna
+            try:
+                sorter = np.lexsort((val, ids))
+            except TypeError:  # catches object dtypes
+                msg = f"val.dtype must be object, got {val.dtype}"
+                assert val.dtype == object, msg
+                val, sorter, _isna = _object_sorter(val, ids)
+            else:
+                _isna = isna
 
         ids, val = ids[sorter], val[sorter]
 

pandas/tests/groupby/test_function.py

@@ -966,6 +966,7 @@ def test_frame_describe_unstacked_format():
 @pytest.mark.parametrize("dropna", [False, True])
 def test_series_groupby_nunique(n, m, sort, dropna):
     def check_nunique(df, keys, as_index=True):
+        original_df = df.copy()
         gr = df.groupby(keys, as_index=as_index, sort=sort)
         left = gr["julie"].nunique(dropna=dropna)
 
@@ -975,6 +976,7 @@ def check_nunique(df, keys, as_index=True):
             right = right.reset_index(drop=True)
 
         tm.assert_series_equal(left, right, check_names=False)
+        tm.assert_frame_equal(df, original_df)
 
     days = date_range("2015-08-23", periods=10)