In February 2026, Alexis Schlomer enhanced the apache/spark repository by implementing Top-K support for the max_by and min_by aggregation functions, introducing three-argument overloads that return arrays of top or bottom k values. This feature was engineered using Scala and Spark’s DataFrame API, leveraging a bounded heap during aggregation to optimize performance and avoid full dataset sorts. Alexis ensured robust test coverage with comprehensive unit tests and a golden SQL file, validating both typical and edge cases. The work streamlined SQL query patterns, reduced reliance on verbose CTEs and window functions, and improved compatibility with platforms like Snowflake and DuckDB.