Data Warehousing

Data warehousing is the practice of collecting data from many sources into a central store designed for analysis and reporting. A data warehouse holds cleaned, modeled, historical data so analysts and BI tools can run complex SQL queries without disturbing the operational systems.

An operational (OLTP) database handles many small reads and writes for running an application, while a data warehouse (OLAP) is organised for scanning and aggregating large volumes of historical data. Warehouses usually load data in batches or streams from operational systems.

A star schema is a dimensional model with a central fact table of measurements, such as sales, linked to dimension tables such as date, product and customer. A snowflake schema normalises the dimensions further. Both are described on the Kimball Group and Wikipedia pages listed here.

Kimball builds dimensional data marts of facts and dimensions. Inmon starts from a normalised, enterprise-wide warehouse that feeds marts. Data Vault models hubs, links and satellites to keep history and integrate many sources. Teams often combine them with layered designs.

ETL transforms data before loading it into the warehouse, while ELT loads raw data first and transforms it inside the warehouse using its compute, which is common with cloud platforms and tools such as dbt.

A warehouse stores modeled data for fast SQL analytics, a data lake stores raw data of any type cheaply in object storage, and a lakehouse adds table formats and catalogs on top of lake storage to give warehouse-like features. See the Data Lakes and Data Lakehouse Platforms pages.

Start with where your data and workloads already live, then compare pricing model, concurrency, latency, open table format support, security, governance and integration with your ETL and BI tools. The Cloud Data Warehouses page lists 26 platforms, and a proof of concept on your own workload is the best test.

Outbound links were checked on 3 November 2026. Snowflake and Microsoft Fabric return 403 to automated checks and are kept, so they may need a manual click. Descriptions follow each publisher's own pages, listing is not an endorsement, and product names and features change often.