Fix IntervalIndex.union to preserve type-metadata (#14051)

Fixes: #14041 

This PR fixes `fillna` that will preserve the type-metadata for `IntervalColumn`.

Authors:
  - GALI PREM SAGAR (https://github.com/galipremsagar)

Approvers:
  - Bradley Dice (https://github.com/bdice)

URL: #14051

python/cudf/cudf/core/column/categorical.py

@@ -1272,18 +1272,7 @@ def fillna(
                     self.codes.dtype
                 )
 
-        result = super().fillna(value=fill_value, method=method)
-
-        result = column.build_categorical_column(
-            categories=self.dtype.categories._values,
-            codes=column.build_column(result.base_data, dtype=result.dtype),
-            offset=result.offset,
-            size=result.size,
-            mask=result.base_mask,
-            ordered=self.dtype.ordered,
-        )
-
-        return result
+        return super().fillna(value=fill_value, method=method)
 
     def indices_of(
         self, value: ScalarLike

python/cudf/cudf/core/column/column.py

@@ -715,7 +715,7 @@ def fillna(
         """
         return libcudf.replace.replace_nulls(
             input_col=self, replacement=value, method=method, dtype=dtype
-        )
+        )._with_type_metadata(self.dtype)
 
     def isnull(self) -> ColumnBase:
         """Identify missing values in a Column."""

python/cudf/cudf/core/column/decimal.py

@@ -147,10 +147,7 @@ def fillna(
                 "integer values"
             )
 
-        result = libcudf.replace.replace_nulls(
-            input_col=self, replacement=value, method=method, dtype=dtype
-        )
-        return result._with_type_metadata(self.dtype)
+        return super().fillna(value=value, method=method)
 
     def normalize_binop_value(self, other):
         if isinstance(other, ColumnBase):

python/cudf/cudf/tests/test_index.py

@@ -2050,6 +2050,10 @@ def test_range_index_concat(objs):
         (pd.Index([0, 1, 2, 30], name="a"), [90, 100]),
         (pd.Index([0, 1, 2, 30]), pd.Index([0, 10, 1.0, 11])),
         (pd.Index(["a", "b", "c", "d", "c"]), pd.Index(["a", "c", "z"])),
+        (
+            pd.IntervalIndex.from_tuples([(0, 2), (0, 2), (2, 4)]),
+            pd.IntervalIndex.from_tuples([(0, 2), (2, 4)]),
+        ),
     ],
 )
 @pytest.mark.parametrize("sort", [None, False])