Is Spark what you're looking for? You can do all sorts of joins, groupings, and aggregations with parquet(s) acting as your source(s)