Data Glossary 🧠
What is ELT?
ELT (Extract, Load, and Transform) is a data integration approach that extracts (E) data from a source system, and loads (L) raw data into a destination system before it transforms (T) the data. In other words, in the ELT approach, transformation (T) of the data is done within the destination Data Warehouse after data has been loaded.
ELT is in contrast to the more traditional ETL data integration approach, in which data is transformed before it arrives at the destination. See ETL vs ELT for a more detailed comparision of these approaches.
The shift from the ETL paradigm to the ELT paradigm has been made possible thanks to the plummeting cost of cloud-based computation and storage, and the appearance of cloud-based data warehouses like Redshift, BigQuery, or Snowflake.
The following image demonstrates the ELT approach to data integration – in this diagram dbt creates and manages the SQL that is used for transforming the data in the destination:
ELT is also related to Reverse ETL which you can find more information about at: Reverse ETL Explained or Airbyte.com.