DBSP: Automatic Incremental View Maintenance for Rich Query Languages

DBSP (Dataflow-Based Signal Processing) is a programming language designed for incremental computation, aiming to efficiently update results by reusing the original output. DBSP employs simple yet powerful operators to handle complex database queries, including relational algebra and streaming computations. It simplifies traditional database management by treating database states as streams and leveraging Z-sets to manage changes. An algorithm is proposed to create circuits that maintain incremental views, promising efficient and modular query processing. The Rust implementation of DBSP provides APIs and a SQL-to-DBSP compiler, demonstrating practical viability and performance.