ETL (Extract, Transform, Load)

ETL — extract, transform, load — is the pipeline pattern in which data is transformed before it reaches the destination: pulled from sources, reshaped and cleaned on separate processing infrastructure, and loaded into the warehouse already finished. Only prepared data lands.

ETL made sense when warehouse compute was scarce and expensive; its costs are that transformation logic lives outside the warehouse (harder for analysts to see or change) and raw data is often discarded, so history can’t be re-derived. Its modern successor, ELT, loads raw data first and transforms inside the warehouse — but ETL survives wherever data must be cleaned, masked, or contracted before it’s allowed to land.

Go deeper: ETL vs ELT: The Same Three Letters in a Different Order