Worked on the apache/phoenix repository to enhance the stability and correctness of the Partitioned CDC Index path, focusing on database and distributed systems reliability. Addressed a specific bug in PartitionIdFunction.java by ensuring that the key is retrieved with tuple.getKey(ptr) before performing length checks, which prevents potential issues when accessing partition IDs for partitioned tables. This targeted fix reduces runtime risks associated with CDC indexing and improves the robustness of handling partitioned datasets. The work was implemented in Java, demonstrating attention to detail in managing data integrity and system reliability within distributed database environments, with a focus on backend correctness.