Data Warehouse Definition by Bill Imnon
A Data Warehouse is a subject-oriented, integrated, non-volatile, and time-variant collection of data in support of management decisions.
Subject-orientation
- Not for one single purpose. One DW for the whole company.
- Subjects: customers, products, activities, sales
Integration
- Data comes from verious sources: Different Databases, SQL, NoSQL, CSV, APIs, …
- Challanges: Data Cleaning (Different sources have different encodings, relations, syntaxes, …), Information Integration, Overcoming Heregogenity
Non-volatile
- Typically, data in a data warehouse is not updated. -> Stable and persistent database
Time-variant
- Historical snapshots of the data with time references. Example for products that change:
| product_id | description | price | manufaturer | valid_from | valid_until |
|---|---|---|---|---|---|
| 17 | Chocolate Bar | 0.89 | Monsterfood | 2023-01-01 | 2023-03-01 |
| 17 | Chocolate Bar | 0.99 | Monsterfood | 2023-03-01 | 9999-12-31 |