BUG: Fix memory leak when slicing Series and assigning to self

pandas/core/series.py

@@ -928,7 +928,7 @@ def _slice(self, slobj: slice, axis: AxisInt = 0) -> Series:
         # axis kwarg is retained for compat with NDFrame method
         #  _slice is *always* positional
         mgr = self._mgr.get_slice(slobj, axis=axis)
-        out = self._constructor_from_mgr(mgr, axes=mgr.axes)
+        out = self._constructor_from_mgr(mgr, axes=mgr.axes).copy(deep=True)
         out._name = self._name
         return out.__finalize__(self)
 

pandas/tests/series/methods/test_slice.py

@@ -0,0 +1,27 @@
+import gc
+
+from pandas import Series
+
+
+class Dummy:
+    def __init__(self, val):
+        self.val = val
+
+
+def count_dummy_instances():
+    gc.collect()
+    return sum(1 for obj in gc.get_objects() if isinstance(obj, Dummy))
+
+
+def test_slicing_releases_dummy_instances():
+    """Ensure Series slicing does not retain references to original Dummy data."""
+    NDATA = 100_000
+    baseline = count_dummy_instances()
+    a = Series([Dummy(i) for i in range(NDATA)])
+    a = a[-1:]
+    gc.collect()
+    after = count_dummy_instances()
+    retained = after - baseline
+    assert retained <= 1, (
+        f"{retained} Dummy instances were retained; expected at most 1"
+    )