// the find
quarylabs/sqruff
Fast SQL formatter/linter
sqruff is a SQL linter and formatter written in Rust, targeting teams who want SQLFluff-style rules but found it too slow. It covers ANSI, BigQuery, ClickHouse, and Snowflake dialects and ships a Python package that adds dbt/Jinja templating on top of the Rust core. Useful if you have a dbt project and are tired of waiting on SQLFluff.
The Rust parser is genuinely fast — the CodSpeed benchmarks are there and the repo takes performance seriously, not just as a marketing claim. The Python bridge for dbt and Jinja templating is a pragmatic call: rather than rewriting an entire templating ecosystem, they call out to Python where necessary and stay in Rust for the parse/lint/format hot path. The WASM playground is a low-friction way to test rules without installing anything. The snapshot-based test suite (`.exitcode`/`.stderr`/`.stdout` files per test case) makes regressions obvious and diff-friendly in PRs.
The dbt/Jinja templating support requires a Python install, which breaks the 'drop a single binary on CI' story — you now have a hybrid runtime dependency that can drift. The `.hacking/dialect_sqlfluff_catchup/` directory exists explicitly to track how far behind SQLFluff they still are on dialect coverage, which is honest but also means any non-mainstream SQL syntax is a gamble. At 1.3k stars versus SQLFluff's 9k+, the community rule library and editor plugin ecosystem are thin; you will hit edge cases with no existing workaround. The project is essentially a single-company bet on Quary Labs continuing to prioritize it — there's no visible contributor base outside that org.