I'm a couple years late to the party but DuckDB is blowing my mind. I couldn't find an good embedded DuckDB stream processing solution so I hacked my own, calling it SQLFlow:<p><a href="https://github.com/turbolytics/sql-flow">https://github.com/turbolytics/sql-flow</a><p>SQLFlow enables writing stream transformations in pure sql, executed using DuckDB.<p>The goal was to create a lightweight, performant stream processing engine using pure SQL transformations, DuckDB didn't disappoint!<p>I'd love your feedback, feature requests, impressions, or just comments. I'd love to turn this into a stable, usable project that people are comfortable running in production.<p>Thank you