Worked on the apache/incubator-gluten repository to enhance the stability of Iceberg partition handling, focusing on improving reliability and data access for partitioned Iceberg tables. Addressed a bug related to partition information retrieval when a partition column is dropped by refactoring partition field filtering logic to include only relevant fields and avoid void transform fields. Added regression tests for v1 Iceberg tables with dropped partitions to ensure ongoing correctness and prevent runtime errors. Utilized Apache Iceberg, Apache Spark, and SQL to implement these changes, providing a more robust foundation for future integrations and aligning the system with data correctness expectations.