DuckDB v2.0 turns the in-process database into a server with Quack and CONNECT
DuckDB's fall v2.0 release graduates the quack extension to stable, so any DuckDB process can serve its databases over the network and another attaches with the new CONNECT statement - the first real break from in-process-only, built on more than 10,000 commits since v1.5 in March. The same release adds BEFORE/AFTER triggers, a PEG-based parser with PostgreSQL dialect compatibility modes, async I/O for object stores, native timezone handling that drops the ICU dependency for a 2-2.6x speedup, and a 40x faster recursive CTE path. Breaking changes stay short - chiefly the new default storage format and the finished lambda-syntax transition - so an existing DuckDB deployment can pick most of this up without a rewrite.
Source: duckdb.org ↗
Where last year was the year of the lakehouse, this release kicks off the year of DuckDB as a server.
Why this matters
- → DuckDB transforms from single-process library into networked database with Quack protocol
- → 10,000+ commits since March add triggers, async I/O, 40x recursive CTE speedup, native timezone handling
- → Breaking changes minimal—existing deployments can adopt most features without rewrite