Chapter 9: Loading Data into a Data Mart

While the data lake enables a significant amount of analytics to happen inside it, there are several use cases where a data engineer may need to load data into an external data warehouse, or data mart, to enable a set of data consumers.

As we reviewed in Chapter 2, Data Management Architectures for Analytics, a data lake is a single source of truth across multiple lines of business, while a data mart contains a subset of data of interest to a particular group of users. A data mart could be a relational database, a data warehouse, or a different kind of data store.

Data marts serve two primary purposes. First, they provide a database with a subset of the data in the data lake, optimized for specific types of queries (such as for a specific business function). In addition, they also provide a higher-performing, lower latency query engine, which is often required for specific analytic use cases (such as for powering business intelligence...