When companies consider the difference between data warehouses vs data marts, they are comparing databases with integrated data vs data silos. Learn more about the differences in our comparative infographic and see why Teradata recommends a unified ecosystem for data storage and data analytics.
What is the Difference Between Data Warehouses vs Data Marts?
A data warehouse is a design pattern and architecture for shared and detailed data. It has many characteristics. Hundreds of sources and applications can send data for all subject areas to the same data warehouse, where it’s integrated and shared across all users.
A data mart is a data repository with a single focus area. It offers a subset of a data warehouse and serves a narrow group of users. Data marts offer only a slice of the overall data in a data warehouse.
What Data do Data Warehouses Offer over Data Marts?
1 single, integrated, connected data warehouse ecosystem that brings together all data, regardless of subject area. This allows one source of truth for all users.
1 data mart per subject area. This can lead to data silos, with each one serving a small group of users.
This can be the difference between 1,000 data silos or 1 data warehouse with the same information.
How Many Tables Do They Typically Have?
> 4,000 tables for a data warehouse
< 20 tables for a data mart
How Much Data Do They Usually Store?
> 100GB data warehouse
< 100GB data mart
What Are the Business and IT Implications?
Data warehouses bring together all data, which is cleansed, managed, integrated, and governed for a complete view of the business.
Data marts create silos, which add more costs, complexity, and data movement. Data marts often have duplicate and redundant data, which results in “data drift,” or data that unexpectedly changes and breaks processes.
What Does Teradata Recommend?
Teradata advocates for making data a company’s greatest asset. This requires breaking down barriers and silos, and aggregating all data for analytics.
Teradata recommends a cloud data analytics platform that unifies everything, including data lakes, data marts, data warehouses, analytics, and new data sources and types. This makes all data searchable and usable. Any user in any department gains the ability to ask any question against any data at any time.