REGR: Fixed hash_key=None for object values (#30900)

* REGR: Fixed hash_key=None for object values

Closes #30887

Author: TomAugspurger

pandas/core/util/hashing.py

@@ -2,6 +2,7 @@
 data hash pandas / numpy objects
 """
 import itertools
+from typing import Optional
 
 import numpy as np
 
@@ -58,7 +59,7 @@ def hash_pandas_object(
     obj,
     index: bool = True,
     encoding: str = "utf8",
-    hash_key: str = _default_hash_key,
+    hash_key: Optional[str] = _default_hash_key,
     categorize: bool = True,
 ):
     """
@@ -82,6 +83,9 @@ def hash_pandas_object(
     """
     from pandas import Series
 
+    if hash_key is None:
+        hash_key = _default_hash_key
+
     if isinstance(obj, ABCMultiIndex):
         return Series(hash_tuples(obj, encoding, hash_key), dtype="uint64", copy=False)
 

pandas/tests/util/test_hashing.py

@@ -374,3 +374,10 @@ def test_hash_with_tuple():
     df3 = pd.DataFrame({"data": [tuple([1, []]), tuple([2, {}])]})
     with pytest.raises(TypeError, match="unhashable type: 'list'"):
         hash_pandas_object(df3)
+
+
+def test_hash_object_none_key():
+    # https://github.com/pandas-dev/pandas/issues/30887
+    result = pd.util.hash_pandas_object(pd.Series(["a", "b"]), hash_key=None)
+    expected = pd.Series([4578374827886788867, 17338122309987883691], dtype="uint64")
+    tm.assert_series_equal(result, expected)