ETL

ETL stands for extract, transform, and load. The term is very widely used in data management and data engineering.

ETL was defined decades ago to describe the process of moving data from transactional systems, such as a billing system, to analytics systems. Data is extracted from the billing system; transformed in various ways, including being converted to analytics-friendly formats such as columnar data stores; then loaded into the analytics system. 

Now ETL is used for multiple steps in data management processes. For instance, in a medallion architecture, ETL is likely to occur between bronze and silver stages, between silver and gold stages, and perhaps at additional points as well. (If ETL is not used at a given step, ELT, defined separately here, is likely to be used instead.)

Reverse ETL takes data that has been processed previously, and has arrived in a data warehouse or data lakehouse, and copies it into business applications such as customer records management (CRM), analytics, and marketing automation tools. This allows these apps to be used more effectively, improving both business results and the quantity and quality of data generated by these applications for use in later processes.

Related terms: ELT; medallion architecture

On the Onehouse website: 

• Apache Hudi vs Delta Lake vs Apache Iceberg
• Getting Started: Incrementally process data with Apache Hudi
• An Introduction to the Hudi and Flink Integration