Commiting progress on default=None. Still not tested

Adding computation.rst section (still not written)

Author: carlosdanielcsantos

doc/source/computation.rst

@@ -459,6 +459,27 @@ default of the index) in a DataFrame.
    dft
    dft.rolling('2s', on='foo').sum()
 
+.. _stats.rolling_window.endpoints:
+
+Rolling window endpoint inclusion
+~~~~~~~~~~~~~~~~~~
+
+.. versionadded:: 0.20.0
+
+New in version 0.19.0 are the ability to pass an offset (or convertible) to a ``.rolling()`` method and have it produce
+variable sized windows based on the passed time window. For each time point, this includes all preceding values occurring
+within the indicated time delta.
+
+This can be particularly useful for a non-regular time frequency index.
+
+.. ipython:: python
+
+   dft = pd.DataFrame({'B': [0, 1, 2, np.nan, 4]},
+                      index=pd.date_range('20130101 09:00:00', periods=5, freq='s'))
+   dft
+
+This is a regular frequency index. Using an integer window parameter works to roll along the window frequency.
+
 .. _stats.moments.ts-versus-resampling:
 
 Time-aware Rolling vs. Resampling

doc/source/whatsnew/v0.20.0.txt

@@ -319,7 +319,7 @@ To convert a ``SparseDataFrame`` back to sparse SciPy matrix in COO format, you
 Other Enhancements
 ^^^^^^^^^^^^^^^^^^
 
-- ``DataFrame.rolling()`` now accepts the parameter ``closed='right'|'left'|'both'|'neither'`` to choose the rolling window endpoint closedness. (:issue:`13965`)
+- ``DataFrame.rolling()`` now accepts the parameter ``closed='right'|'left'|'both'|'neither'`` to choose the rolling window endpoint closedness. See :ref:`documentation <stats.rolling_window.endpoints>` (:issue:`13965`)
 - Integration with the ``feather-format``, including a new top-level ``pd.read_feather()`` and ``DataFrame.to_feather()`` method, see :ref:`here <io.feather>`.
 - ``Series.str.replace()`` now accepts a callable, as replacement, which is passed to ``re.sub`` (:issue:`15055`)
 - ``Series.str.replace()`` now accepts a compiled regular expression as a pattern (:issue:`15446`)

pandas/core/window.py

@@ -60,7 +60,7 @@ class _Window(PandasObject, SelectionMixin):
     exclusions = set()
 
     def __init__(self, obj, window=None, min_periods=None, freq=None,
-                 center=False, win_type=None, axis=0, on=None, closed='right',
+                 center=False, win_type=None, axis=0, on=None, closed=None,
                  **kwargs):
 
         if freq is not None:
@@ -379,10 +379,12 @@ class Window(_Window):
     on : string, optional
         For a DataFrame, column on which to calculate
         the rolling window, rather than the index
-    closed : 'right', 'left', 'both', 'neither'
-        For offset-based windows, make the interval closed on the right, left,
-        or on both endpoints. Can also make the interval open on both endpoints
-        (neither).
+    closed : string, default None
+        Make the interval closed on the 'right', 'left', 'both' or
+        'neither' endpoints.
+        For offset-based windows, it defaults to 'right'.
+        For fixed windows, defaults to 'both'. Remaining cases not implemented
+        for fixed windows.
 
         .. versionadded:: 0.20.0
 

pandas/core/window.pyx

@@ -165,8 +165,8 @@ cdef class MockFixedWindowIndexer(WindowIndexer):
 
     """
     def __init__(self, ndarray input, int64_t win, int64_t minp,
-                 object index=None, object floor=None,
-                 bint left_closed=False, bint right_closed=True):
+                 bint left_closed, bint right_closed,
+                 object index=None, object floor=None):
 
         assert index is None
         self.is_variable = 0
@@ -203,8 +203,8 @@ cdef class FixedWindowIndexer(WindowIndexer):
 
     """
     def __init__(self, ndarray input, int64_t win, int64_t minp,
-                 object index=None, object floor=None,
-                 bint left_closed=False, bint right_closed=True):
+                 bint left_closed, bint right_closed,
+                 object index=None, object floor=None):
         cdef ndarray start_s, start_e, end_s, end_e
 
         assert index is None
@@ -352,7 +352,11 @@ def get_window_indexer(input, win, minp, index, closed,
         bint left_closed = False
         bint right_closed = False
 
-    assert closed in ['right', 'left', 'both', 'neither']
+    assert closed is None or closed in ['right', 'left', 'both', 'neither']
+
+    # if windows is variable, default is 'right', otherwise default is 'both'
+    if closed is None:
+        closed = 'right' if index is not None else 'both'
 
     if closed in ['right', 'both']:
         right_closed = True

pandas/tests/test_window.py

@@ -3405,6 +3405,10 @@ def test_closed(self):
         result = df.rolling('2s', closed='right').sum()
         tm.assert_frame_equal(result, expected)
 
+        # default should be 'right'
+        result = df.rolling('2s').sum()
+        tm.assert_frame_equal(result, expected)
+
         expected = df.copy()
         expected["A"] = [1.0, 2, 3, 3, 2]
         result = df.rolling('2s', closed='both').sum()