dbt (Data Build Tool)
dbt is a popular tool for the “T” (transform) step inside a warehouse: you write your transformations as plain SQL SELECT statements, and dbt turns them into managed tables and views, in the right order, with tests and documentation. It is the “T” in ELT done with software-engineering discipline.
dbt caught on because it lets analysts treat data transformations like real code: each model is version-controlled in Git, models can reference each other so dbt automatically figures out the build order (a dependency graph), and you can attach data tests (“this id column must be unique and never null”) that fail loudly when assumptions break. The result is transformations that are readable, reusable, tested, and documented — instead of a tangle of one-off SQL scripts.
dbt does not move data; it only transforms data already loaded inside your warehouse.