Fix index caching.

Author: DriesSchaumont

pandas/_libs/reduction.pyx

@@ -55,7 +55,7 @@ cdef class _BaseGrouper:
 
     cdef inline _update_cached_objs(self, object cached_typ, object cached_ityp,
                                     Slider islider, Slider vslider):
-        if cached_typ is None or len(vslider.buf) != len(cached_ityp):
+        if cached_typ is None:
             cached_ityp = self.ityp(islider.buf)
             cached_typ = self.typ(
                 vslider.buf, dtype=vslider.buf.dtype, index=cached_ityp, name=self.name
@@ -65,12 +65,12 @@ cdef class _BaseGrouper:
             # We need this for EA-backed indexes that have a reference
             # to a 1-d ndarray like datetime / timedelta / period.
             object.__setattr__(cached_ityp, '_index_data', islider.buf)
+            object.__setattr__(cached_ityp, '_data', islider.buf)
             cached_ityp._engine.clear_mapping()
             cached_ityp._cache.clear()  # e.g. inferred_freq must go
             cached_typ._mgr.set_values(vslider.buf)
             object.__setattr__(cached_typ, '_index', cached_ityp)
             object.__setattr__(cached_typ, 'name', self.name)
-
         return cached_typ, cached_ityp
 
     cdef inline object _apply_to_group(self,
@@ -259,6 +259,7 @@ cdef class SeriesGrouper(_BaseGrouper):
                         start += group_size
                         group_size = 0
                         continue
+                    print(f"Group size: {group_size}")
 
                     end = start + group_size
                     islider.move(start, end)

pandas/tests/groupby/test_bin_groupby.py

@@ -27,6 +27,21 @@ def test_series_grouper():
     tm.assert_almost_equal(counts, exp_counts)
 
 
+def test_series_grouper_result_length_difference():
+    obj = Series(np.random.randn(10), dtype="float64")
+    obj.index = obj.index.astype("O")
+    labels = np.array([-1, -1, -1, 0, 0, 0, 1, 1, 1, 1], dtype=np.intp)
+
+    grouper = libreduction.SeriesGrouper(obj, lambda x: all(x > 0), labels, 2)
+    result, counts = grouper.get_result()
+
+    expected = np.array([all(obj[3:6] > 0), all(obj[6:] > 0)])
+    tm.assert_almost_equal(result, expected)
+
+    exp_counts = np.array([3, 4], dtype=np.int64)
+    tm.assert_almost_equal(counts, exp_counts)
+
+
 def test_series_grouper_requires_nonempty_raises():
     # GH#29500
     obj = Series(np.random.randn(10))