Fix groupby agg preserve subclass type

doc/source/whatsnew/v2.3.0.rst

@@ -142,7 +142,7 @@ Plotting
 
 Groupby/resample/rolling
 ^^^^^^^^^^^^^^^^^^^^^^^^
--
+- Bug in :meth:`DataFrame.groupby` followed by :meth:`DataFrameGroupBy.agg` not preserving subclass type of the original DataFrame (:issue:`59667`)
 -
 
 Reshaping

pandas/core/groupby/generic.py

@@ -1585,6 +1585,9 @@ def aggregate(self, func=None, *args, engine=None, engine_kwargs=None, **kwargs)
             result = self._insert_inaxis_grouper(result)
             result.index = default_index(len(result))
 
+        if isinstance(self.obj, type(self.obj)):
+            result = self.obj._constructor(result)
+
         return result
 
     agg = aggregate

pandas/tests/groupby/test_groupby.py

@@ -3005,6 +3005,36 @@ def test_groupby_agg_namedagg_with_duplicate_columns():
     tm.assert_frame_equal(result, expected)
 
 
+class MyDataFrame(DataFrame):
+    @property
+    def _constructor(self):
+        return MyDataFrame
+
+
+@pytest.mark.parametrize(
+    "data, agg_dict, expected",
+    [
+        pytest.param(
+            {"A": [1, 1, 2, 2], "B": [1, 2, 3, 4]},
+            {"B": "sum"},
+            DataFrame({"B": [3, 7]}, index=Index([1, 2], name="A")),
+        ),
+        pytest.param(
+            {"A": [1, 1, 2, 2], "B": [1, 2, 3, 4], "C": [4, 3, 2, 1]},
+            {"B": "sum", "C": "mean"},
+            DataFrame({"B": [3, 7], "C": [3.5, 1.5]}, index=Index([1, 2], name="A")),
+        ),
+    ],
+)
+def test_groupby_agg_preserves_subclass(data, agg_dict, expected):
+    # GH#59667
+    df = MyDataFrame(data)
+    result = df.groupby("A").agg(agg_dict)
+
+    assert isinstance(result, MyDataFrame)
+    tm.assert_frame_equal(result, expected)
+
+
 def test_groupby_multi_index_codes():
     # GH#54347
     df = DataFrame(