Is ETL dead?
ETL stands for Extract, Transform, Load. It is an orchestrated “three-step” process that extracts Data from one or more data sources, applies necessary transformations (cleaning, merging, sorting, filtering…) and loads the data into a target system [1].
ETL is a foundational data integration pattern for enterprise integrations. It’s used every day by a huge number of businesses. But many of those processes were built in the objective of feeding Data Warehouses, in some cases, many years ago.
As we moved with real-time integration pattern, a drop in using ETL was observed, in favor of real-time integrations, which made the question “Is ETL dead” legitimate. The target systems also evolved, from having Data Warehouses, to other Data platforms, like Data Lakes, or Data Lakehouses implying major evolution in the data integration patterns.
So, is ETL dead and can businesses shut down the doors to their data warehouses?
What are ETL steps?
EXTRACT: The process extracts data from one or more data sources.
TRANSFORM: The incoming data goes through a transformation layer to make the data suitable for the target system and its data model.
LOAD: It consists of loading the transformed extracted data into the target system [1].
