There are two distinct schools of thought on how to build a data warehouse: the Inmon approach and the Kimball approach. Their key difference is how data structures are modeled, loaded and stored in a data warehouse. This difference affects the delivery time of the data warehouse and the ability to accommodate future changes in the ETL design.
When a data architect is asked to design and implement a data warehouse from scratch, which architectural style should he or she choose to build a data warehouse? How to help architects choose between Inmon or Kimball architecture?
Inmon's three-paradigm modeling is often mobile number list compared with Kimball's dimensional data, and the two great gods have always adhered to their own data modeling views. Two great gods had a very interesting point, Kimball once said: "A data warehouse is nothing but a combination of all data marts", to which Inmon responded: "You can catch all the little fish in the ocean and They get together - but they still can't be whales."
In a typical data warehouse, we start with a set of OLTP data sources (for the description of OLTP, see These can be Excel sheets, ERP systems, documents or basically any other data source. After the data is stored in the target environment, the data is processed and transformed using ETL tools and then fed into the data warehouse.
Inmon believes that data should go directly into the data warehouse after the ETL process. Kimball insists that after the ETL process, the data should be loaded into data marts, and the union of all these data marts creates a conceptual data warehouse.