Bringing Efficiency and Applicability to User-Defined Functions by WebAssembly

With the continuous evolution of database technology, numerous databases and query analysis engines have begun to incorporate user-defined functions into SQL. These functions enable users to express their unique data processing logic. However, Python, being the most commonly used language for implementing such functions, faces certain performance and applicability challenges. In this paper, we propose the utilization of WebAssembly for implementing user-defined functions. WebAssembly offers significant performance advantages over Python and supports compilation from multiple languages, thereby allowing users to leverage their preferred programming languages for writing user-defined functions. This approach enhances the overall applicability of user-defined functions. Furthermore, we introduce several performance optimization strategies for WebAssembly-based user-defined functions. These include leveraging a sophisticated user-defined functions execution engine, employing JIT compilation, and utilizing vectorization technology. These optimizations contribute significantly to enhancing the performance of WebAssembly-based user-defined functions.