Worked on the apache/spark repository to enhance the reliability of DataFrame operations by addressing a specific issue in Spark SQL resolution. Focused on correcting the handling of UnresolvedAttribute within the DataFrameDropColumns logic, the work prevented incorrect attribute propagation from grandchild outputs, which previously risked producing inaccurate query results and downstream pipeline failures. The solution involved a targeted bug fix, carefully aligned with SPARK-49782 conventions for traceability and code quality. Utilizing Scala and Spark, the developer applied data analysis expertise to improve the correctness and stability of DataFrame drop operations, ultimately reducing runtime errors and supporting more robust data pipelines.