Worked on the xupefei/delta repository, focusing on improving the Spark Delta Connector by addressing a critical metadata issue. Using Scala and leveraging expertise in Delta Lake and Spark, corrected the PARQUET_FIELD_NESTED_IDS_METADATA_KEY within the COLUMN_MAPPING_METADATA_KEYS to prevent erroneous metadata-driven mappings. The approach involved updating the relevant code and enhancing unit test coverage to ensure the fix was robust and regression-proof. No user-facing features were introduced, but the work reduced the risk of downstream ingestion failures and contributed to the overall stability of the integration. The changes prepared the codebase for safer and more maintainable future development.