During the month, contributed to the apache/paimon repository by addressing a critical bug in the PaimonMetadataApplier component. Focused on backend development and data engineering, the work involved refining how partition keys are managed within the Flink CDC pipeline. By ensuring that partition keys are only appended to primary keys when primary keys exist, the solution stabilized metadata handling for append-only partitioning scenarios. This adjustment, implemented in Java and leveraging Apache Flink, reduced the risk of metadata drift and ingestion errors, thereby improving data correctness and reliability for downstream analytics. The contribution reflects careful attention to data pipeline integrity and maintainability.