🧠 Second Brain


Search IconIcon to open search


Last updated Feb 15, 2024

ETL, which stands for Extract, Transform, and Load, involves a three-phase process for moving data. Unlike ELT, ETL requires pre-defining the data schema and completing transformations before the data reaches its final destination. In contrast, ELT involves loading data first and then transforming it.

# Evolution of ETL

ETL processes are evolving. Historically, they were executed using tools like Informatica, IBM Datastage, Cognos, AbInitio, or Microsoft SSIS. Nowadays, the trend leans towards more programmatic or configuration-driven platforms such as Apache Airflow, Dagster, and Temporal. This shift coincides with growing data demands and the need for quicker data accessibility, steering the trend towards ELT.

# ETL vs ELT

ETL (Extract, Transform, Load) and ELT (Extract, Load, Transform) are two methodologies for data handling. Traditionally, ETL has been the backbone of Data Warehouse processes, while ELT is more commonly associated with Data Lake creation.

See also: ETL is changing, ETL with Apache Airflow.

# ETL Tools

See ETL Tools.

References: ETL vs ELT, ELT