fix for pandas-dev#60695 fix Series constructor dropping key levels when keys have varying entry counts

Author: mansoor17syed

pandas/core/indexes/multi.py

@@ -526,58 +526,36 @@ def from_tuples(
     ) -> MultiIndex:
         """
         Convert list of tuples to MultiIndex.
-
-        Parameters
-        ----------
-        tuples : list / sequence of tuple-likes
-            Each tuple is the index of one row/column.
-        sortorder : int or None
-            Level of sortedness (must be lexicographically sorted by that
-            level).
-        names : list / sequence of str, optional
-            Names for the levels in the index.
-
-        Returns
-        -------
-        MultiIndex
         """
         if not is_list_like(tuples):
             raise TypeError("Input must be a list / sequence of tuple-likes.")
+        
         if is_iterator(tuples):
             tuples = list(tuples)
+        
+        # Cast to proper tuple type
         tuples = cast(Collection[tuple[Hashable, ...]], tuples)
-
-        # handling the empty tuple cases
-        if len(tuples) and all(isinstance(e, tuple) and not e for e in tuples):
-            codes = [np.zeros(len(tuples))]
-            levels = [Index(com.asarray_tuplesafe(tuples, dtype=np.dtype("object")))]
-            return cls(
-                levels=levels,
-                codes=codes,
-                sortorder=sortorder,
-                names=names,
-                verify_integrity=False,
-            )
-
-        arrays: list[Sequence[Hashable]]
+        
         if len(tuples) == 0:
             if names is None:
                 raise TypeError("Cannot infer number of levels from empty list")
-            # error: Argument 1 to "len" has incompatible type "Hashable";
-            # expected "Sized"
-            arrays = [[]] * len(names)  # type: ignore[arg-type]
+            arrays = [[]] * len(names)
         elif isinstance(tuples, (np.ndarray, Index)):
             if isinstance(tuples, Index):
                 tuples = np.asarray(tuples._values)
-
             arrays = list(lib.tuples_to_object_array(tuples).T)
         elif isinstance(tuples, list):
-            arrays = list(lib.to_object_array_tuples(tuples).T)
+            # Find the maximum length of tuples
+            max_length = max((len(t) for t in tuples), default=0)
+            
+            # Pad shorter tuples with np.nan
+            padded_tuples = [
+                t + (np.nan,) * (max_length - len(t)) for t in tuples
+            ]
+            arrays = list(lib.to_object_array_tuples(padded_tuples).T)
         else:
-            # Use zip_longest instead of zip to handle tuples of different lengths
             from itertools import zip_longest
-            arrs = zip_longest(*tuples, fillvalue=np.nan)
-            arrays = cast(list[Sequence[Hashable]], arrs)
+            arrays = list(zip_longest(*tuples, fillvalue=np.nan))
 
         return cls.from_arrays(arrays, sortorder=sortorder, names=names)
 

pandas/tests/indexes/multi/test_constructors.py

@@ -865,47 +865,43 @@ def test_from_tuples_different_lengths_gh60695():
     
     GH#60695
     """
-    # Test case 1: Original issue example
+    # Test case 1: Original issue example with string values
     tuples = [("l1",), ("l1", "l2")]
     result = pd.MultiIndex.from_tuples(tuples)
     expected = pd.MultiIndex.from_tuples([("l1", np.nan), ("l1", "l2")])
-    tm.assert_index_equal(result, expected)
+    tm.assert_index_equal(result, expected, exact=True)
 
     # Test case 2: Series construction with tuple keys
     s = pd.Series({("l1",): "v1", ("l1", "l2"): "v2"})
     expected = pd.Series(
         ["v1", "v2"],
         index=pd.MultiIndex.from_tuples([("l1", np.nan), ("l1", "l2")])
     )
-    tm.assert_series_equal(s, expected)
+    tm.assert_series_equal(s, expected, check_index_type=True)
 
-    # Test case 3: Multiple levels with None
-    data = {(1, 1, None): -1.0}
-    result = pd.Series(data)
-    expected = pd.Series(
-        -1.0,
-        index=pd.MultiIndex.from_tuples([(1, 1, np.nan)]),
-    )
-    tm.assert_series_equal(result, expected)
+    # Test case 3: Handle numeric values
+    tuples = [(1,), (1, 2)]
+    result = pd.MultiIndex.from_tuples(tuples)
+    expected = pd.MultiIndex.from_tuples([(1, np.nan), (1, 2)])
+    tm.assert_index_equal(result, expected, exact=True)
 
-    # Test case 4: Mixed length tuples
-    tuples = [("a",), ("b", "c"), ("d", "e", "f")]
+    # Test case 4: Mixed types (strings and integers)
+    tuples = [(1, "a"), (1,), (2, "b", "c")]
     result = pd.MultiIndex.from_tuples(tuples)
     expected = pd.MultiIndex.from_tuples([
-        ("a", np.nan, np.nan),
-        ("b", "c", np.nan),
-        ("d", "e", "f")
+        (1, "a", np.nan),
+        (1, np.nan, np.nan),
+        (2, "b", "c")
     ])
-    tm.assert_index_equal(result, expected)
+    tm.assert_index_equal(result, expected, exact=True)
 
-    # Test case 5: DataFrame with tuple index
-    df = pd.DataFrame(
-        {"col": ["v1", "v2"]},
-        index=pd.MultiIndex.from_tuples([("l1",), ("l1", "l2")])
-    )
-    expected_index = pd.MultiIndex.from_tuples([("l1", np.nan), ("l1", "l2")])
-    expected_df = pd.DataFrame(
-        {"col": ["v1", "v2"]},
-        index=expected_index
-    )
-    tm.assert_frame_equal(df, expected_df)
+    # Test case 5: Empty tuples
+    tuples = []
+    with pytest.raises(TypeError, match="Cannot infer number of levels"):
+        pd.MultiIndex.from_tuples(tuples)
+
+    # Test case 6: Single level consistency
+    tuples = [("a",)]
+    result = pd.MultiIndex.from_tuples(tuples)
+    expected = pd.MultiIndex.from_tuples([("a",)])
+    tm.assert_index_equal(result, expected, exact=True)