extract, transform, load (ETL)

An ETL process is the data flow process into a data warehouse or related system. In this process, data is extracted from the source systems, transformed, integrated and written to the target system.

The ETL process (extract/transform/load) characterizes the three classic steps for filling a data warehouse:

1. Extraction: data is loaded from one or more source systems into the ETL system.


ETL process: Extraction, transformation, loading

2. Transformation: The data is transformed in the ETL system. This can involve technical transformations, such as changing data types, as well as business transformations, such as calculations and aggregations.
3. Load: The data is loaded from the ETL system into the data warehouse.

In modern ETL processes, a distinction is made between other tasks that are performed in an ETL process:

1. data cleansing (data cleansing)

2. quality assurance

3. historization

4. pre- aggregation

5. cube and datamart generation

6. master data management

7. email notification, etc.

8. logging and change tracking.