From 861f762e9cba7d411f95c279ab25c7343fc9961d Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Wed, 10 Feb 2021 11:03:53 -0800
Subject: [PATCH 1/9] Remove _groupby object from class

---
 pandas/core/groupby/groupby.py  | 30 ++++++++++++--
 pandas/core/window/ewm.py       |  2 +-
 pandas/core/window/expanding.py |  2 +-
 pandas/core/window/rolling.py   | 72 +++++++++++++++++++++------------
 4 files changed, 75 insertions(+), 31 deletions(-)

diff --git a/pandas/core/groupby/groupby.py b/pandas/core/groupby/groupby.py
index 5758762c13984..3c6f8e0d5cda2 100644
--- a/pandas/core/groupby/groupby.py
+++ b/pandas/core/groupby/groupby.py
@@ -1883,7 +1883,15 @@ def rolling(self, *args, **kwargs):
         """
         from pandas.core.window import RollingGroupby
 
-        return RollingGroupby(self, *args, **kwargs)
+        return RollingGroupby(
+            self._selected_obj,
+            *args,
+            _grouping_indices=self.grouper.indices,
+            _grouping_keys=self.grouper.names,
+            _grouping_codes=self.grouper.codes,
+            _grouping_levels=self.grouper.levels,
+            **kwargs,
+        )
 
     @final
     @Substitution(name="groupby")
@@ -1895,7 +1903,15 @@ def expanding(self, *args, **kwargs):
         """
         from pandas.core.window import ExpandingGroupby
 
-        return ExpandingGroupby(self, *args, **kwargs)
+        return ExpandingGroupby(
+            self._selected_obj,
+            *args,
+            _grouping_indices=self.grouper.indices,
+            _grouping_keys=self.grouper.names,
+            _grouping_codes=self.grouper.codes,
+            _grouping_levels=self.grouper.levels,
+            **kwargs,
+        )
 
     @final
     @Substitution(name="groupby")
@@ -1906,7 +1922,15 @@ def ewm(self, *args, **kwargs):
         """
         from pandas.core.window import ExponentialMovingWindowGroupby
 
-        return ExponentialMovingWindowGroupby(self, *args, **kwargs)
+        return ExponentialMovingWindowGroupby(
+            self._selected_obj,
+            *args,
+            _grouping_indices=self.grouper.indices,
+            _grouping_keys=self.grouper.names,
+            _grouping_codes=self.grouper.codes,
+            _grouping_levels=self.grouper.levels,
+            **kwargs,
+        )
 
     @final
     def _fill(self, direction, limit=None):
diff --git a/pandas/core/window/ewm.py b/pandas/core/window/ewm.py
index 633427369902d..13ecc0acffd28 100644
--- a/pandas/core/window/ewm.py
+++ b/pandas/core/window/ewm.py
@@ -548,7 +548,7 @@ class ExponentialMovingWindowGroupby(BaseWindowGroupby, ExponentialMovingWindow)
 
     @property
     def _constructor(self):
-        return ExponentialMovingWindow
+        return ExponentialMovingWindowGroupby
 
     def _get_window_indexer(self) -> GroupbyIndexer:
         """
diff --git a/pandas/core/window/expanding.py b/pandas/core/window/expanding.py
index f91441de41448..eb3d8c482d80d 100644
--- a/pandas/core/window/expanding.py
+++ b/pandas/core/window/expanding.py
@@ -617,7 +617,7 @@ class ExpandingGroupby(BaseWindowGroupby, Expanding):
 
     @property
     def _constructor(self):
-        return Expanding
+        return ExpandingGroupby
 
     def _get_window_indexer(self) -> GroupbyIndexer:
         """
diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index bec6cfb375716..6dfa1f4dcced5 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -4,6 +4,7 @@
 """
 from __future__ import annotations
 
+import copy
 from datetime import timedelta
 from functools import partial
 import inspect
@@ -53,7 +54,7 @@
 from pandas.core.base import DataError, SelectionMixin
 import pandas.core.common as common
 from pandas.core.construction import extract_array
-from pandas.core.groupby.base import GotItemMixin, ShallowMixin
+from pandas.core.groupby.base import ShallowMixin
 from pandas.core.indexes.api import Index, MultiIndex
 from pandas.core.reshape.concat import concat
 from pandas.core.util.numba_ import NUMBA_FUNC_CACHE, maybe_use_numba
@@ -518,19 +519,40 @@ def aggregate(self, func, *args, **kwargs):
     agg = aggregate
 
 
-class BaseWindowGroupby(GotItemMixin, BaseWindow):
+class BaseWindowGroupby(BaseWindow):
     """
     Provide the groupby windowing facilities.
     """
 
-    def __init__(self, obj, *args, **kwargs):
-        kwargs.pop("parent", None)
-        groupby = kwargs.pop("groupby", None)
-        if groupby is None:
-            groupby, obj = obj, obj._selected_obj
-        self._groupby = groupby
-        self._groupby.mutated = True
-        self._groupby.grouper.mutated = True
+    _attributes: List[str] = [
+        "window",
+        "min_periods",
+        "center",
+        "win_type",
+        "axis",
+        "on",
+        "closed",
+        "method",
+        "_grouping_indices",
+        "_grouping_keys",
+        "_grouping_codes",
+        "_grouping_levels",
+    ]
+
+    def __init__(
+        self,
+        obj,
+        *args,
+        _grouping_indices=None,
+        _grouping_keys=None,
+        _grouping_codes=None,
+        _grouping_levels=None,
+        **kwargs,
+    ):
+        self._grouping_indices = _grouping_indices or {}
+        self._grouping_keys = _grouping_keys or []
+        self._grouping_codes = _grouping_codes or []
+        self._grouping_levels = _grouping_levels or []
         super().__init__(obj, *args, **kwargs)
 
     def _apply(
@@ -550,9 +572,7 @@ def _apply(
         # 1st set of levels = group by labels
         # 2nd set of levels = original index
         # Ignore 2nd set of levels if a group by label include an index level
-        result_index_names = [
-            grouping.name for grouping in self._groupby.grouper._groupings
-        ]
+        result_index_names = copy.copy(self._grouping_keys)
         grouped_object_index = None
 
         column_keys = [
@@ -569,10 +589,10 @@ def _apply(
             # Our result will have still kept the column in the result
             result = result.drop(columns=column_keys, errors="ignore")
 
-        codes = self._groupby.grouper.codes
-        levels = self._groupby.grouper.levels
+        codes = self._grouping_codes
+        levels = copy.copy(self._grouping_levels)
 
-        group_indices = self._groupby.grouper.indices.values()
+        group_indices = self._grouping_indices.values()
         if group_indices:
             indexer = np.concatenate(list(group_indices))
         else:
@@ -606,7 +626,7 @@ def _apply_pairwise(
         Apply the given pairwise function given 2 pandas objects (DataFrame/Series)
         """
         # Manually drop the grouping column first
-        target = target.drop(columns=self._groupby.grouper.names, errors="ignore")
+        target = target.drop(columns=self._grouping_keys, errors="ignore")
         result = super()._apply_pairwise(target, other, pairwise, func)
         # 1) Determine the levels + codes of the groupby levels
         if other is not None:
@@ -617,12 +637,12 @@ def _apply_pairwise(
             result = concat(
                 [
                     result.take(gb_indices).reindex(result.index)
-                    for gb_indices in self._groupby.indices.values()
+                    for gb_indices in self._grouping_indices.values()
                 ]
             )
 
             gb_pairs = (
-                common.maybe_make_list(pair) for pair in self._groupby.indices.keys()
+                common.maybe_make_list(pair) for pair in self._grouping_indices.keys()
             )
             groupby_codes = []
             groupby_levels = []
@@ -636,10 +656,10 @@ def _apply_pairwise(
         else:
             # When we evaluate the pairwise=True result, repeat the groupby
             # labels by the number of columns in the original object
-            groupby_codes = self._groupby.grouper.codes
-            groupby_levels = self._groupby.grouper.levels
+            groupby_codes = self._grouping_codes
+            groupby_levels = self._grouping_levels
 
-            group_indices = self._groupby.grouper.indices.values()
+            group_indices = self._grouping_indices.values()
             if group_indices:
                 indexer = np.concatenate(list(group_indices))
             else:
@@ -666,7 +686,7 @@ def _apply_pairwise(
         # 3) Create the resulting index by combining 1) + 2)
         result_codes = groupby_codes + result_codes
         result_levels = groupby_levels + result_levels
-        result_names = self._groupby.grouper.names + result_names
+        result_names = self._grouping_keys + result_names
 
         result_index = MultiIndex(
             result_levels, result_codes, names=result_names, verify_integrity=False
@@ -683,7 +703,7 @@ def _create_data(self, obj: FrameOrSeries) -> FrameOrSeries:
         # GH 36197
         if not obj.empty:
             groupby_order = np.concatenate(
-                list(self._groupby.grouper.indices.values())
+                list(self._grouping_indices.values())
             ).astype(np.int64)
             obj = obj.take(groupby_order)
         return super()._create_data(obj)
@@ -2144,7 +2164,7 @@ class RollingGroupby(BaseWindowGroupby, Rolling):
 
     @property
     def _constructor(self):
-        return Rolling
+        return RollingGroupby
 
     def _get_window_indexer(self) -> GroupbyIndexer:
         """
@@ -2174,7 +2194,7 @@ def _get_window_indexer(self) -> GroupbyIndexer:
         window_indexer = GroupbyIndexer(
             index_array=index_array,
             window_size=window,
-            groupby_indicies=self._groupby.indices,
+            groupby_indicies=self._grouping_indices,
             window_indexer=rolling_indexer,
             indexer_kwargs=indexer_kwargs,
         )

From 0543909d4aa254e21f3fdccd34023c50a07ba13d Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Wed, 10 Feb 2021 23:36:27 -0800
Subject: [PATCH 2/9] Add test that groupby apply is not affected

---
 pandas/core/window/ewm.py           |  2 +-
 pandas/core/window/expanding.py     |  2 +-
 pandas/tests/window/test_groupby.py | 34 +++++++++++++++++++++++++++++
 3 files changed, 36 insertions(+), 2 deletions(-)

diff --git a/pandas/core/window/ewm.py b/pandas/core/window/ewm.py
index 13ecc0acffd28..5c0fea7152b04 100644
--- a/pandas/core/window/ewm.py
+++ b/pandas/core/window/ewm.py
@@ -559,7 +559,7 @@ def _get_window_indexer(self) -> GroupbyIndexer:
         GroupbyIndexer
         """
         window_indexer = GroupbyIndexer(
-            groupby_indicies=self._groupby.indices,
+            groupby_indicies=self._grouping_indices,
             window_indexer=ExponentialMovingWindowIndexer,
         )
         return window_indexer
diff --git a/pandas/core/window/expanding.py b/pandas/core/window/expanding.py
index eb3d8c482d80d..8888a19e38d9d 100644
--- a/pandas/core/window/expanding.py
+++ b/pandas/core/window/expanding.py
@@ -628,7 +628,7 @@ def _get_window_indexer(self) -> GroupbyIndexer:
         GroupbyIndexer
         """
         window_indexer = GroupbyIndexer(
-            groupby_indicies=self._groupby.indices,
+            groupby_indicies=self._grouping_indices,
             window_indexer=ExpandingIndexer,
         )
         return window_indexer
diff --git a/pandas/tests/window/test_groupby.py b/pandas/tests/window/test_groupby.py
index d3c2b5467e5bb..df31d24cc5326 100644
--- a/pandas/tests/window/test_groupby.py
+++ b/pandas/tests/window/test_groupby.py
@@ -82,6 +82,9 @@ def test_rolling(self, f):
         r = g.rolling(window=4)
 
         result = getattr(r, f)()
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.rolling(4), f)())
         tm.assert_frame_equal(result, expected)
 
@@ -91,6 +94,9 @@ def test_rolling_ddof(self, f):
         r = g.rolling(window=4)
 
         result = getattr(r, f)(ddof=1)
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.rolling(4), f)(ddof=1))
         tm.assert_frame_equal(result, expected)
 
@@ -101,6 +107,9 @@ def test_rolling_quantile(self, interpolation):
         g = self.frame.groupby("A")
         r = g.rolling(window=4)
         result = r.quantile(0.4, interpolation=interpolation)
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(
             lambda x: x.rolling(4).quantile(0.4, interpolation=interpolation)
         )
@@ -136,6 +145,9 @@ def test_rolling_apply(self, raw):
 
         # reduction
         result = r.apply(lambda x: x.sum(), raw=raw)
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(lambda x: x.rolling(4).apply(lambda y: y.sum(), raw=raw))
         tm.assert_frame_equal(result, expected)
 
@@ -643,6 +655,16 @@ def test_groupby_rolling_resulting_multiindex(self):
         )
         tm.assert_index_equal(result.index, expected_index)
 
+    def test_groupby_rolling_object_doesnt_affect_groupby_apply(self):
+        # GH 39732
+        g = self.frame.groupby("A")
+        expected = g.apply(lambda x: x.rolling(4).sum()).index
+        _ = g.rolling(window=4)
+        result = g.apply(lambda x: x.rolling(4).sum()).index
+        tm.assert_index_equal(result, expected)
+        assert not g.mutated
+        assert not g.grouper.mutated
+
 
 class TestExpanding:
     def setup_method(self):
@@ -656,6 +678,9 @@ def test_expanding(self, f):
         r = g.expanding()
 
         result = getattr(r, f)()
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.expanding(), f)())
         tm.assert_frame_equal(result, expected)
 
@@ -665,6 +690,9 @@ def test_expanding_ddof(self, f):
         r = g.expanding()
 
         result = getattr(r, f)(ddof=0)
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.expanding(), f)(ddof=0))
         tm.assert_frame_equal(result, expected)
 
@@ -675,6 +703,9 @@ def test_expanding_quantile(self, interpolation):
         g = self.frame.groupby("A")
         r = g.expanding()
         result = r.quantile(0.4, interpolation=interpolation)
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(
             lambda x: x.expanding().quantile(0.4, interpolation=interpolation)
         )
@@ -714,6 +745,9 @@ def test_expanding_apply(self, raw):
 
         # reduction
         result = r.apply(lambda x: x.sum(), raw=raw)
+        # GH 39732
+        g.mutated = True
+        g.grouper.mutated = True
         expected = g.apply(lambda x: x.expanding().apply(lambda y: y.sum(), raw=raw))
         tm.assert_frame_equal(result, expected)
 

From 7e572bec8f3378020ed69a4d7cd44ad1415a1ada Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Thu, 11 Feb 2021 10:43:09 -0800
Subject: [PATCH 3/9] Remove unnecessary groupby subclasses

---
 pandas/core/groupby/base.py   | 11 ++---------
 pandas/core/window/rolling.py | 11 +++++++----
 2 files changed, 9 insertions(+), 13 deletions(-)

diff --git a/pandas/core/groupby/base.py b/pandas/core/groupby/base.py
index 594c5899209df..99426c55da29b 100644
--- a/pandas/core/groupby/base.py
+++ b/pandas/core/groupby/base.py
@@ -5,7 +5,6 @@
 """
 import collections
 from typing import List
-import warnings
 
 from pandas._typing import final
 
@@ -28,10 +27,7 @@ def _shallow_copy(self, obj, **kwargs):
             obj = obj.obj
         for attr in self._attributes:
             if attr not in kwargs:
-                # TODO: Remove once win_type deprecation is enforced
-                with warnings.catch_warnings():
-                    warnings.filterwarnings("ignore", "win_type", FutureWarning)
-                    kwargs[attr] = getattr(self, attr)
+                kwargs[attr] = getattr(self, attr)
         return self._constructor(obj, **kwargs)
 
 
@@ -63,10 +59,7 @@ def _gotitem(self, key, ndim, subset=None):
 
         # we need to make a shallow copy of ourselves
         # with the same groupby
-        # TODO: Remove once win_type deprecation is enforced
-        with warnings.catch_warnings():
-            warnings.filterwarnings("ignore", "win_type", FutureWarning)
-            kwargs = {attr: getattr(self, attr) for attr in self._attributes}
+        kwargs = {attr: getattr(self, attr) for attr in self._attributes}
 
         # Try to select from a DataFrame, falling back to a Series
         try:
diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index 6dfa1f4dcced5..e832db47fe1d3 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -54,7 +54,6 @@
 from pandas.core.base import DataError, SelectionMixin
 import pandas.core.common as common
 from pandas.core.construction import extract_array
-from pandas.core.groupby.base import ShallowMixin
 from pandas.core.indexes.api import Index, MultiIndex
 from pandas.core.reshape.concat import concat
 from pandas.core.util.numba_ import NUMBA_FUNC_CACHE, maybe_use_numba
@@ -89,7 +88,7 @@
     from pandas.core.internals import Block  # noqa:F401
 
 
-class BaseWindow(ShallowMixin, SelectionMixin):
+class BaseWindow(SelectionMixin):
     """Provides utilities for performing windowing operations."""
 
     _attributes: List[str] = [
@@ -237,8 +236,12 @@ def _gotitem(self, key, ndim, subset=None):
         # create a new object to prevent aliasing
         if subset is None:
             subset = self.obj
-        self = self._shallow_copy(subset)
-        self._reset_cache()
+        # TODO: Remove once win_type deprecation is enforced
+        with warnings.catch_warnings():
+            warnings.filterwarnings("ignore", "win_type", FutureWarning)
+            self = type(self)(
+                subset, **{attr: getattr(self, attr) for attr in self._attributes}
+            )
         if subset.ndim == 2:
             if is_scalar(key) and key in subset or is_list_like(key):
                 self._selection = key

From ef3f79144541d7ab942dea7135b65de82c1c2b09 Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Thu, 11 Feb 2021 10:58:42 -0800
Subject: [PATCH 4/9] Supress private variables in repr

---
 pandas/core/window/rolling.py | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index e832db47fe1d3..4475dc6e59b1e 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -116,8 +116,6 @@ def __init__(
         method: str = "single",
         **kwargs,
     ):
-
-        self.__dict__.update(kwargs)
         self.obj = obj
         self.on = on
         self.closed = closed
@@ -267,7 +265,7 @@ def __repr__(self) -> str:
         attrs_list = (
             f"{attr_name}={getattr(self, attr_name)}"
             for attr_name in self._attributes
-            if getattr(self, attr_name, None) is not None
+            if getattr(self, attr_name, None) is not None and attr_name[0] != "_"
         )
         attrs = ",".join(attrs_list)
         return f"{type(self).__name__} [{attrs}]"

From 3d76221a14c8ad9cec2d4719a32d21ce22582b1e Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Thu, 11 Feb 2021 11:38:20 -0800
Subject: [PATCH 5/9] Don't have window classes accept kwargs

---
 pandas/core/window/ewm.py             | 29 ++++++++++++-----
 pandas/core/window/expanding.py       |  6 ++--
 pandas/core/window/rolling.py         | 47 +++++++++++++++------------
 pandas/tests/window/test_expanding.py |  2 +-
 4 files changed, 51 insertions(+), 33 deletions(-)

diff --git a/pandas/core/window/ewm.py b/pandas/core/window/ewm.py
index 5c0fea7152b04..55f68254bb4f6 100644
--- a/pandas/core/window/ewm.py
+++ b/pandas/core/window/ewm.py
@@ -203,7 +203,15 @@ class ExponentialMovingWindow(BaseWindow):
     4  3.233686
     """
 
-    _attributes = ["com", "min_periods", "adjust", "ignore_na", "axis"]
+    _attributes = [
+        "com",
+        "min_periods",
+        "adjust",
+        "ignore_na",
+        "axis",
+        "halflife",
+        "times",
+    ]
 
     def __init__(
         self,
@@ -217,17 +225,18 @@ def __init__(
         ignore_na: bool = False,
         axis: int = 0,
         times: Optional[Union[str, np.ndarray, FrameOrSeries]] = None,
-        **kwargs,
     ):
-        self.obj = obj
-        self.min_periods = max(int(min_periods), 1)
+        super().__init__(
+            obj=obj,
+            min_periods=max(int(min_periods), 1),
+            on=None,
+            center=False,
+            closed=None,
+            method="single",
+            axis=axis,
+        )
         self.adjust = adjust
         self.ignore_na = ignore_na
-        self.axis = axis
-        self.on = None
-        self.center = False
-        self.closed = None
-        self.method = "single"
         if times is not None:
             if isinstance(times, str):
                 times = self._selected_obj[times]
@@ -546,6 +555,8 @@ class ExponentialMovingWindowGroupby(BaseWindowGroupby, ExponentialMovingWindow)
     Provide an exponential moving window groupby implementation.
     """
 
+    _attributes = ExponentialMovingWindow._attributes + BaseWindowGroupby._attributes
+
     @property
     def _constructor(self):
         return ExponentialMovingWindowGroupby
diff --git a/pandas/core/window/expanding.py b/pandas/core/window/expanding.py
index 8888a19e38d9d..c825ca833431d 100644
--- a/pandas/core/window/expanding.py
+++ b/pandas/core/window/expanding.py
@@ -80,9 +80,7 @@ class Expanding(RollingAndExpandingMixin):
 
     _attributes = ["min_periods", "center", "axis", "method"]
 
-    def __init__(
-        self, obj, min_periods=1, center=None, axis=0, method="single", **kwargs
-    ):
+    def __init__(self, obj, min_periods=1, center=None, axis=0, method="single"):
         super().__init__(
             obj=obj, min_periods=min_periods, center=center, axis=axis, method=method
         )
@@ -615,6 +613,8 @@ class ExpandingGroupby(BaseWindowGroupby, Expanding):
     Provide a expanding groupby implementation.
     """
 
+    _attributes = Expanding._attributes + BaseWindowGroupby._attributes
+
     @property
     def _constructor(self):
         return ExpandingGroupby
diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index 4475dc6e59b1e..4e523eb4bb5f1 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -91,16 +91,7 @@
 class BaseWindow(SelectionMixin):
     """Provides utilities for performing windowing operations."""
 
-    _attributes: List[str] = [
-        "window",
-        "min_periods",
-        "center",
-        "win_type",
-        "axis",
-        "on",
-        "closed",
-        "method",
-    ]
+    _attributes: List[str] = []
     exclusions: Set[str] = set()
 
     def __init__(
@@ -114,7 +105,6 @@ def __init__(
         on: Optional[Union[str, Index]] = None,
         closed: Optional[str] = None,
         method: str = "single",
-        **kwargs,
     ):
         self.obj = obj
         self.on = on
@@ -525,15 +515,7 @@ class BaseWindowGroupby(BaseWindow):
     Provide the groupby windowing facilities.
     """
 
-    _attributes: List[str] = [
-        "window",
-        "min_periods",
-        "center",
-        "win_type",
-        "axis",
-        "on",
-        "closed",
-        "method",
+    _attributes = [
         "_grouping_indices",
         "_grouping_keys",
         "_grouping_codes",
@@ -895,6 +877,17 @@ class Window(BaseWindow):
     2013-01-01 09:00:06  4.0
     """
 
+    _attributes = [
+        "window",
+        "min_periods",
+        "center",
+        "win_type",
+        "axis",
+        "on",
+        "closed",
+        "method",
+    ]
+
     def validate(self):
         super().validate()
 
@@ -1385,6 +1378,18 @@ def corr_func(x, y):
 
 
 class Rolling(RollingAndExpandingMixin):
+
+    _attributes = [
+        "window",
+        "min_periods",
+        "center",
+        "win_type",
+        "axis",
+        "on",
+        "closed",
+        "method",
+    ]
+
     def validate(self):
         super().validate()
 
@@ -2163,6 +2168,8 @@ class RollingGroupby(BaseWindowGroupby, Rolling):
     Provide a rolling groupby implementation.
     """
 
+    _attributes = Rolling._attributes + BaseWindowGroupby._attributes
+
     @property
     def _constructor(self):
         return RollingGroupby
diff --git a/pandas/tests/window/test_expanding.py b/pandas/tests/window/test_expanding.py
index 01804faad5a5e..554b8a404f714 100644
--- a/pandas/tests/window/test_expanding.py
+++ b/pandas/tests/window/test_expanding.py
@@ -49,7 +49,7 @@ def test_constructor_invalid(frame_or_series, w):
 @pytest.mark.parametrize("method", ["std", "mean", "sum", "max", "min", "var"])
 def test_numpy_compat(method):
     # see gh-12811
-    e = Expanding(Series([2, 4, 6]), window=2)
+    e = Expanding(Series([2, 4, 6]))
 
     msg = "numpy operations are not valid with window objects"
 

From f914c665d4a38f87bb3bc1c8d1ddafd338bc933a Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Thu, 11 Feb 2021 17:07:34 -0800
Subject: [PATCH 6/9] Add whatsnew

---
 doc/source/whatsnew/v1.3.0.rst | 1 +
 1 file changed, 1 insertion(+)

diff --git a/doc/source/whatsnew/v1.3.0.rst b/doc/source/whatsnew/v1.3.0.rst
index 09e1853429d9f..2242acb642acb 100644
--- a/doc/source/whatsnew/v1.3.0.rst
+++ b/doc/source/whatsnew/v1.3.0.rst
@@ -416,6 +416,7 @@ Groupby/resample/rolling
 - Bug in :meth:`Series.resample` would raise when the index was a :class:`PeriodIndex` consisting of ``NaT`` (:issue:`39227`)
 - Bug in :meth:`core.window.rolling.RollingGroupby.corr` and :meth:`core.window.expanding.ExpandingGroupby.corr` where the groupby column would return 0 instead of ``np.nan`` when providing ``other`` that was longer than each group (:issue:`39591`)
 - Bug in :meth:`core.window.expanding.ExpandingGroupby.corr` and :meth:`core.window.expanding.ExpandingGroupby.cov` where 1 would be returned instead of ``np.nan`` when providing ``other`` that was longer than each group (:issue:`39591`)
+- Bug in :meth:`DataFrameGroupBy.apply` where a :class:`MultiIndex` would be created instead of an :class:`Index` if a :class:`:meth:`core.window.rolling.RollingGroupby` object was created (:issue:`39732`)
 
 Reshaping
 ^^^^^^^^^

From f330a78f0d62cf4571d4792551be15054c924309 Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Thu, 11 Feb 2021 17:10:59 -0800
Subject: [PATCH 7/9] Remove unused constructor

---
 pandas/core/window/ewm.py       | 4 ----
 pandas/core/window/expanding.py | 4 ----
 pandas/core/window/rolling.py   | 4 ----
 3 files changed, 12 deletions(-)

diff --git a/pandas/core/window/ewm.py b/pandas/core/window/ewm.py
index 55f68254bb4f6..4c757e5e91671 100644
--- a/pandas/core/window/ewm.py
+++ b/pandas/core/window/ewm.py
@@ -557,10 +557,6 @@ class ExponentialMovingWindowGroupby(BaseWindowGroupby, ExponentialMovingWindow)
 
     _attributes = ExponentialMovingWindow._attributes + BaseWindowGroupby._attributes
 
-    @property
-    def _constructor(self):
-        return ExponentialMovingWindowGroupby
-
     def _get_window_indexer(self) -> GroupbyIndexer:
         """
         Return an indexer class that will compute the window start and end bounds
diff --git a/pandas/core/window/expanding.py b/pandas/core/window/expanding.py
index c825ca833431d..b13484c80c82b 100644
--- a/pandas/core/window/expanding.py
+++ b/pandas/core/window/expanding.py
@@ -615,10 +615,6 @@ class ExpandingGroupby(BaseWindowGroupby, Expanding):
 
     _attributes = Expanding._attributes + BaseWindowGroupby._attributes
 
-    @property
-    def _constructor(self):
-        return ExpandingGroupby
-
     def _get_window_indexer(self) -> GroupbyIndexer:
         """
         Return an indexer class that will compute the window start and end bounds
diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index 8a0e813e2351e..9f5b20d6ae54b 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -2170,10 +2170,6 @@ class RollingGroupby(BaseWindowGroupby, Rolling):
 
     _attributes = Rolling._attributes + BaseWindowGroupby._attributes
 
-    @property
-    def _constructor(self):
-        return RollingGroupby
-
     def _get_window_indexer(self) -> GroupbyIndexer:
         """
         Return an indexer class that will compute the window start and end bounds

From 885745a773a36e338f21b7f1a6b389fd9ece0cd8 Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Mon, 22 Feb 2021 20:23:30 -0800
Subject: [PATCH 8/9] Just user the grouper

---
 pandas/core/groupby/groupby.py  | 15 ++--------
 pandas/core/window/ewm.py       |  2 +-
 pandas/core/window/expanding.py |  2 +-
 pandas/core/window/rolling.py   | 49 ++++++++++++++-------------------
 4 files changed, 25 insertions(+), 43 deletions(-)

diff --git a/pandas/core/groupby/groupby.py b/pandas/core/groupby/groupby.py
index 76802e6c4b53b..8e9cf066d39e2 100644
--- a/pandas/core/groupby/groupby.py
+++ b/pandas/core/groupby/groupby.py
@@ -1921,10 +1921,7 @@ def rolling(self, *args, **kwargs):
         return RollingGroupby(
             self._selected_obj,
             *args,
-            _grouping_indices=self.grouper.indices,
-            _grouping_keys=self.grouper.names,
-            _grouping_codes=self.grouper.codes,
-            _grouping_levels=self.grouper.levels,
+            _grouper=self.grouper,
             **kwargs,
         )
 
@@ -1941,10 +1938,7 @@ def expanding(self, *args, **kwargs):
         return ExpandingGroupby(
             self._selected_obj,
             *args,
-            _grouping_indices=self.grouper.indices,
-            _grouping_keys=self.grouper.names,
-            _grouping_codes=self.grouper.codes,
-            _grouping_levels=self.grouper.levels,
+            _grouper=self.grouper,
             **kwargs,
         )
 
@@ -1960,10 +1954,7 @@ def ewm(self, *args, **kwargs):
         return ExponentialMovingWindowGroupby(
             self._selected_obj,
             *args,
-            _grouping_indices=self.grouper.indices,
-            _grouping_keys=self.grouper.names,
-            _grouping_codes=self.grouper.codes,
-            _grouping_levels=self.grouper.levels,
+            _grouper=self.grouper,
             **kwargs,
         )
 
diff --git a/pandas/core/window/ewm.py b/pandas/core/window/ewm.py
index 034c1d96a0aa6..208b5ab0023eb 100644
--- a/pandas/core/window/ewm.py
+++ b/pandas/core/window/ewm.py
@@ -576,7 +576,7 @@ def _get_window_indexer(self) -> GroupbyIndexer:
         GroupbyIndexer
         """
         window_indexer = GroupbyIndexer(
-            groupby_indicies=self._grouping_indices,
+            groupby_indicies=self._grouper.indices,
             window_indexer=ExponentialMovingWindowIndexer,
         )
         return window_indexer
diff --git a/pandas/core/window/expanding.py b/pandas/core/window/expanding.py
index 3a87ad19fd286..c201216a91ab1 100644
--- a/pandas/core/window/expanding.py
+++ b/pandas/core/window/expanding.py
@@ -638,7 +638,7 @@ def _get_window_indexer(self) -> GroupbyIndexer:
         GroupbyIndexer
         """
         window_indexer = GroupbyIndexer(
-            groupby_indicies=self._grouping_indices,
+            groupby_indicies=self._grouper.indices,
             window_indexer=ExpandingIndexer,
         )
         return window_indexer
diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index af564dc787b31..844f04ab7c196 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -538,27 +538,18 @@ class BaseWindowGroupby(BaseWindow):
     Provide the groupby windowing facilities.
     """
 
-    _attributes = [
-        "_grouping_indices",
-        "_grouping_keys",
-        "_grouping_codes",
-        "_grouping_levels",
-    ]
+    _attributes = ["_grouper"]
 
     def __init__(
         self,
         obj,
         *args,
-        _grouping_indices=None,
-        _grouping_keys=None,
-        _grouping_codes=None,
-        _grouping_levels=None,
+        _grouper=None,
         **kwargs,
     ):
-        self._grouping_indices = _grouping_indices or {}
-        self._grouping_keys = _grouping_keys or []
-        self._grouping_codes = _grouping_codes or []
-        self._grouping_levels = _grouping_levels or []
+        if _grouper is None:
+            raise ValueError("Must pass a Grouper object.")
+        self._grouper = _grouper
         super().__init__(obj, *args, **kwargs)
 
     def _apply(
@@ -578,7 +569,7 @@ def _apply(
         # 1st set of levels = group by labels
         # 2nd set of levels = original index
         # Ignore 2nd set of levels if a group by label include an index level
-        result_index_names = copy.copy(self._grouping_keys)
+        result_index_names = copy.copy(self._grouper.names)
         grouped_object_index = None
 
         column_keys = [
@@ -595,10 +586,10 @@ def _apply(
             # Our result will have still kept the column in the result
             result = result.drop(columns=column_keys, errors="ignore")
 
-        codes = self._grouping_codes
-        levels = copy.copy(self._grouping_levels)
+        codes = self._grouper.codes
+        levels = copy.copy(self._grouper.levels)
 
-        group_indices = self._grouping_indices.values()
+        group_indices = self._grouper.indices.values()
         if group_indices:
             indexer = np.concatenate(list(group_indices))
         else:
@@ -632,7 +623,7 @@ def _apply_pairwise(
         Apply the given pairwise function given 2 pandas objects (DataFrame/Series)
         """
         # Manually drop the grouping column first
-        target = target.drop(columns=self._grouping_keys, errors="ignore")
+        target = target.drop(columns=self._grouper.names, errors="ignore")
         result = super()._apply_pairwise(target, other, pairwise, func)
         # 1) Determine the levels + codes of the groupby levels
         if other is not None:
@@ -643,12 +634,12 @@ def _apply_pairwise(
             result = concat(
                 [
                     result.take(gb_indices).reindex(result.index)
-                    for gb_indices in self._grouping_indices.values()
+                    for gb_indices in self._grouper.indices.values()
                 ]
             )
 
             gb_pairs = (
-                common.maybe_make_list(pair) for pair in self._grouping_indices.keys()
+                common.maybe_make_list(pair) for pair in self._grouper.indices.keys()
             )
             groupby_codes = []
             groupby_levels = []
@@ -662,10 +653,10 @@ def _apply_pairwise(
         else:
             # When we evaluate the pairwise=True result, repeat the groupby
             # labels by the number of columns in the original object
-            groupby_codes = self._grouping_codes
-            groupby_levels = self._grouping_levels
+            groupby_codes = self._grouper.codes
+            groupby_levels = self._grouper.levels
 
-            group_indices = self._grouping_indices.values()
+            group_indices = self._grouper.indices.values()
             if group_indices:
                 indexer = np.concatenate(list(group_indices))
             else:
@@ -692,7 +683,7 @@ def _apply_pairwise(
         # 3) Create the resulting index by combining 1) + 2)
         result_codes = groupby_codes + result_codes
         result_levels = groupby_levels + result_levels
-        result_names = self._grouping_keys + result_names
+        result_names = self._grouper.names + result_names
 
         result_index = MultiIndex(
             result_levels, result_codes, names=result_names, verify_integrity=False
@@ -708,9 +699,9 @@ def _create_data(self, obj: FrameOrSeries) -> FrameOrSeries:
         # to the groups
         # GH 36197
         if not obj.empty:
-            groupby_order = np.concatenate(
-                list(self._grouping_indices.values())
-            ).astype(np.int64)
+            groupby_order = np.concatenate(list(self._grouper.indices.values())).astype(
+                np.int64
+            )
             obj = obj.take(groupby_order)
         return super()._create_data(obj)
 
@@ -2221,7 +2212,7 @@ def _get_window_indexer(self) -> GroupbyIndexer:
         window_indexer = GroupbyIndexer(
             index_array=index_array,
             window_size=window,
-            groupby_indicies=self._grouping_indices,
+            groupby_indicies=self._grouper.indices,
             window_indexer=rolling_indexer,
             indexer_kwargs=indexer_kwargs,
         )

From 54927a70ec148f9e7ff4dd1c688a837892e575fd Mon Sep 17 00:00:00 2001
From: Matthew Roeschke <emailformattr@gmail.com>
Date: Mon, 22 Feb 2021 20:54:26 -0800
Subject: [PATCH 9/9] Make expected index instead

---
 pandas/tests/window/test_groupby.py | 50 +++++++++++++++--------------
 1 file changed, 26 insertions(+), 24 deletions(-)

diff --git a/pandas/tests/window/test_groupby.py b/pandas/tests/window/test_groupby.py
index df31d24cc5326..c3c5bbe460134 100644
--- a/pandas/tests/window/test_groupby.py
+++ b/pandas/tests/window/test_groupby.py
@@ -82,10 +82,10 @@ def test_rolling(self, f):
         r = g.rolling(window=4)
 
         result = getattr(r, f)()
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.rolling(4), f)())
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize("f", ["std", "var"])
@@ -94,10 +94,10 @@ def test_rolling_ddof(self, f):
         r = g.rolling(window=4)
 
         result = getattr(r, f)(ddof=1)
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.rolling(4), f)(ddof=1))
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize(
@@ -106,13 +106,14 @@ def test_rolling_ddof(self, f):
     def test_rolling_quantile(self, interpolation):
         g = self.frame.groupby("A")
         r = g.rolling(window=4)
+
         result = r.quantile(0.4, interpolation=interpolation)
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(
             lambda x: x.rolling(4).quantile(0.4, interpolation=interpolation)
         )
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize("f", ["corr", "cov"])
@@ -145,10 +146,10 @@ def test_rolling_apply(self, raw):
 
         # reduction
         result = r.apply(lambda x: x.sum(), raw=raw)
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(lambda x: x.rolling(4).apply(lambda y: y.sum(), raw=raw))
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     def test_rolling_apply_mutability(self):
@@ -678,10 +679,10 @@ def test_expanding(self, f):
         r = g.expanding()
 
         result = getattr(r, f)()
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.expanding(), f)())
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize("f", ["std", "var"])
@@ -690,10 +691,10 @@ def test_expanding_ddof(self, f):
         r = g.expanding()
 
         result = getattr(r, f)(ddof=0)
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(lambda x: getattr(x.expanding(), f)(ddof=0))
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize(
@@ -702,13 +703,14 @@ def test_expanding_ddof(self, f):
     def test_expanding_quantile(self, interpolation):
         g = self.frame.groupby("A")
         r = g.expanding()
+
         result = r.quantile(0.4, interpolation=interpolation)
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(
             lambda x: x.expanding().quantile(0.4, interpolation=interpolation)
         )
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize("f", ["corr", "cov"])
@@ -745,10 +747,10 @@ def test_expanding_apply(self, raw):
 
         # reduction
         result = r.apply(lambda x: x.sum(), raw=raw)
-        # GH 39732
-        g.mutated = True
-        g.grouper.mutated = True
         expected = g.apply(lambda x: x.expanding().apply(lambda y: y.sum(), raw=raw))
+        # GH 39732
+        expected_index = MultiIndex.from_arrays([self.frame["A"], range(40)])
+        expected.index = expected_index
         tm.assert_frame_equal(result, expected)