Data warehouse

A data warehouse is a centralized repository that stores structured, cleaned data optimized for reporting and analysis, as opposed to the operational databases that run day-to-day transactions. It typically holds historical data organized into subject areas such as sales, customers or finance, and is queried heavily by BI tools and analysts.

Data arrives through ETL or ELT pipelines, which extract it from source systems and transform it into structured schemas, often a star schema design, before loading it into the warehouse. This differs from a data lake, which accepts raw, often unstructured data in its native format and defers transformation until query time. Warehouses enforce schema up front, which makes them fast for known, repeatable queries but less flexible for exploratory or semi-structured data.

Warehouses are the backbone of most business intelligence programs because they give analysts a consistent, well-modeled source of truth for dashboards and reports. Common pitfalls include letting schemas ossify as the business changes, duplicating transformation logic across reports instead of centralizing it, and underestimating the ongoing cost of data modeling and pipeline maintenance.