In Fabric data pipelines, how do ETL and ELT differ?

• A. ETL loads raw data and performs transformations inside the lakehouse.
• B. ETL is only used for batch processing, not streaming.
• C. ETL transforms data before loading into the lakehouse; ELT loads raw data and performs transformations within the lakehouse using compute engines.
• D. ELT requires dedicated ETL engines outside the lakehouse.

Answer: (C)

Transforming data happens at different points in the pipeline. In ETL, you clean and shape the data before it’s loaded into the lakehouse, so what lands in the store is already transformed. In ELT, you first load the raw data into the lakehouse and then perform the transformations inside the lakehouse using its compute engines. This makes ELT flexible: you retain the raw data and can run different transformations later without re-ingesting. In Fabric, the distinction is between doing the work in an external ETL step versus leveraging the lakehouse’s compute (like Spark or SQL) to transform after loading. The other options misstate where transformations occur or the capabilities of ETL/ELT.