diff --git a/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py b/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py
index 851bed08c9d..3b6bb9b5e83 100644
--- a/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py
+++ b/ingestion/src/metadata/ingestion/source/database/bigquery/metadata.py
@@ -1020,13 +1020,3 @@ class BigquerySource(LifeCycleQueryMixin, CommonDbSourceService, MultiDBSource):
                 )
         else:
             yield from super().mark_tables_as_deleted()
-
-    def _process_col_type(self, column: dict, schema: str) -> Tuple:
-        """
-        Override the parent method to always return parsed_string as None for BigQuery.
-        This ensures we always use the standard column type handling path.
-        """
-        data_type_display, arr_data_type, _ = super()._process_col_type(column, schema)
-        # For BigQuery, we always want to force parsed_string to None
-        # This ensures we use the standard column type handling path
-        return data_type_display, arr_data_type, None
diff --git a/ingestion/src/metadata/ingestion/source/database/sql_column_handler.py b/ingestion/src/metadata/ingestion/source/database/sql_column_handler.py
index 049ae4ff2f4..9f2b8689840 100644
--- a/ingestion/src/metadata/ingestion/source/database/sql_column_handler.py
+++ b/ingestion/src/metadata/ingestion/source/database/sql_column_handler.py
@@ -299,12 +299,6 @@ class SqlColumnHandlerMixin:
                     arrayDataType=arr_data_type,
                     ordinalPosition=column.get("ordinalPosition"),
                 )
-                if column.get("children"):
-                    om_column.children = [
-                        process_column(children) for children in column.get("children")
-                    ]
-                    if not arr_data_type:
-                        om_column.arrayDataType = DataType.UNKNOWN.value
                 if precision:
                     # Precision and scale must be integer values
                     om_column.precision = int(precision[0])
@@ -314,6 +308,15 @@ class SqlColumnHandlerMixin:
                     column=column, parsed_string=parsed_string
                 )
                 om_column = col_obj
+
+                if column.get("children"):
+                    # Prioritize source-provided children for column processing.
+                    # If 'children' are directly provided in the source metadata,
+                    # process and assign them to the output column, overriding any derived children.
+                    # Currently, this is only used for BigQuery.
+                    om_column.children = [
+                        process_column(children) for children in column.get("children")
+                    ]
             om_column.tags = self.get_column_tag_labels(
                 table_name=table_name, column=column
             )