// the find
zzzeek/sqlalchemy
THIS IS NOT THE OFFICIAL REPO - PLEASE SUBMIT PRs ETC AT: http://github.com/sqlalchemy/sqlalchemy
This is zzzeek's personal mirror of SQLAlchemy — the canonical repo is at sqlalchemy/sqlalchemy. SQLAlchemy itself needs no introduction: it's the dominant Python ORM and SQL toolkit, used everywhere from small Flask apps to large-scale data pipelines. This mirror exists because zzzeek is the original author and still the primary maintainer.
The Core/ORM separation is genuinely well-designed — you can use the SQL expression layer without the ORM, which is rare and useful. Bound parameters are used unconditionally, so SQL injection is not something you have to think about. The identity map and unit-of-work implementation handle complex object graphs correctly in ways that ActiveRecord-style ORMs quietly get wrong. The async support added in 2.0 is a real implementation, not a thin wrapper that blocks under load.
This is a mirror repo with 3.4k stars versus the real repo's 10k+; PRs and issues filed here go nowhere, which is a footgun for anyone who lands here via a search. The 2.0 migration from the legacy Query API was breaking enough that a large fraction of tutorials and Stack Overflow answers are now wrong, and there's no fast way to tell which era a given code snippet belongs to. The session/transaction model is powerful but has a steep learning curve — lazy loading, detached instances, and expired attributes bite people regularly in async contexts specifically.