While using the function set_index with the parameter inplace=True, the function attempted to create a new index with a dtype of FLS

S{value}. This dtype was not recognized by the function _dtype_to_subclass, which raised a NotImplementedError. To address this, I added
a verification to the function asarray_tuplesafe that converts data to an array with object type, allowing the index to be created succes
sfully.

Additionally, I created a new test and simplified a previously created test. I also reverted the test file test_parquet.py to restore
the intended FLS behavior.

Author: PF2100

doc/source/whatsnew/v3.0.0.rst

@@ -378,7 +378,7 @@ Performance improvements
 
 Bug fixes
 ~~~~~~~~~
-- Fixed bug in :class:`Index` Index constructor did not allow FLS as indices. (:issue:`57645`)
+- Fixed bug in :class:`Index` Index constructor was not converting FLS to object. (:issue:`57645`)
 - Fixed bug in :class:`SparseDtype` for equal comparison with na fill value. (:issue:`54770`)
 - Fixed bug in :meth:`DataFrame.join` inconsistently setting result index name (:issue:`55815`)
 - Fixed bug in :meth:`DataFrame.to_string` that raised ``StopIteration`` with nested DataFrames. (:issue:`16098`)

pandas/core/common.py

@@ -236,7 +236,7 @@ def asarray_tuplesafe(values: Iterable, dtype: NpDtype | None = None) -> ArrayLi
         values = list(values)
     elif isinstance(values, ABCIndex):
         return values._values
-    elif isinstance(values, ABCSeries):
+    elif isinstance(values, ABCSeries) and values.dtype != "S":
         return values._values
 
     if isinstance(values, list) and dtype in [np.object_, object]:

pandas/core/indexes/base.py

@@ -625,9 +625,7 @@ def _dtype_to_subclass(cls, dtype: DtypeObj):
             # NB: assuming away MultiIndex
             return Index
 
-        elif (
-            dtype.kind == "S" or issubclass(dtype.type, str) or is_numeric_dtype(dtype)
-        ):
+        elif issubclass(dtype.type, str) or is_numeric_dtype(dtype):
             return Index
 
         raise NotImplementedError(dtype)

pandas/tests/frame/methods/test_set_index.py

@@ -618,17 +618,13 @@ def test_set_index_raise_on_len(
             df.set_index(["A", df.A, box(values)], drop=drop, append=append)
 
     def test_set_index_with_FLS_Dtype(self):
-        string_length = 6
-        in_dtype, df_name = f"S{string_length}", "fruit"
-        data = ["apple", "banana", "orange", "grape"]
-
         # Create array with FLS(|S{value}) dtype
-        arr = np.array(data, dtype=in_dtype)
-        df = DataFrame(Series(arr), columns=[df_name])
+        arr = np.array(["apple", "banana", "orange", "grape"], dtype="S6")
+        df = DataFrame(Series(arr), columns=["fruits"])
 
         # This will create a new Index with FLS dtype
-        expected = Index(data=Series(arr), name=df_name)
-        df.set_index(df_name, inplace=True)
+        expected = Index(data=Series(arr), name="fruits")
+        df.set_index("fruits", inplace=True)
         tm.assert_index_equal(df.index, expected)
 
 

pandas/tests/indexes/test_index_new.py

@@ -185,6 +185,18 @@ def test_constructor_datetimes_mixed_tzs(self):
         expected = Index([dt1, dt2], dtype=object)
         tm.assert_index_equal(result, expected)
 
+    def test_FLS_to_object_conversion(self):
+        # Create NumPy array of fixed-length strings
+        arr = np.array(["apple", "banana", "orange", "grape"], dtype="S6")
+        # Create expected array for index
+        expected_arr = np.array(
+            [b"apple", b"banana", b"orange", b"grape"], dtype=object
+        )
+        # This will create a new Index with FLS dtype
+        index = Index(data=Series(arr), name="fruits")
+        expected = Index(data=Series(expected_arr), name="fruits")
+        tm.assert_index_equal(index, expected)
+
 
 class TestDtypeEnforced:
     # check we don't silently ignore the dtype keyword

pandas/tests/io/test_parquet.py

@@ -1027,7 +1027,9 @@ def test_columns_dtypes_not_invalid(self, pa):
 
         # bytes
         df.columns = [b"foo", b"bar"]
-        check_round_trip(df, pa)
+        with pytest.raises(NotImplementedError, match="|S3"):
+            # Bytes fails on read_parquet
+            check_round_trip(df, pa)
 
         # python object
         df.columns = [