ETL

ETL (Extract, Transform, Load) is the pipeline pattern for moving data from source systems into a warehouse: extract from the sources, transform (clean, join, reshape) it, then load it into the target. The newer ELT loads raw data first and transforms inside the warehouse.

Worked example: a nightly ETL pulls orders from Postgres and events from a queue, joins and cleans them, and loads daily aggregates into the warehouse for analysts. Gotcha: ETL transforms before loading (less raw data stored, but reprocessing means re-running the whole pipeline), while ELT loads raw then transforms with the warehouse’s compute (more flexible, more storage) — modern cloud warehouses made ELT the common default.