Data Warehouse

A data warehouse is a system that organizes data for analytical queries and reporting, commonly integrating records from several sources and retaining history. Its design supports questions across many records, such as sales by region over three years.

Different work from taking an order

An order application must record a customer's purchase and update its status. An analyst may need to scan millions of purchases, join them with products, and compare monthly totals. Separating these workloads can prevent a large analytical query from competing directly with checkout activity.

In a warehouse, an order fact table might store measurable transactions, while product and calendar dimension tables describe them. This is one common model, not a requirement that every warehouse use the same table design.

Loading data does not settle its meaning

For a regional sales report, decide whether a refund changes the original sale month or appears in the refund month. Also define currency conversion and whether cancelled orders count. The warehouse stores and queries the chosen representation; it does not choose the correct policy for the business.

AWS describes warehouses in terms of analytical and reporting workloads. Their role is broader than a restriction to one file type. A data lake can also hold structured tables, and a warehouse may support semi-structured data.

A lakehouse brings warehouse-like management and query capabilities to lake storage. The practical comparison is whether the system meets the required performance, governance, and workload needs, rather than whether its name belongs to one exclusive category.

Reference: AWS: Data warehouses.