Mostly abbreviated as DW, DWH or EDW. DW and DWH are used for Data Warehouse and EDW is used for Enterprise Data Warehouse.
Many people have defined data warehouse differently. But also it is said that the most popular definitions came from Bill Inmon and Ralph Kimball.
How Bill Inmon has defined Data Warehouse is as under.
A data warehouse is a subject-oriented, integrated, time-variant and non-volatile collection of data in support of management decision making process.
And according to Ralph Kimball
A data warehouse is a copy of transaction data specifically structured for query and analysis.
This is a functional view of a data warehouse. Ralph Kimball did not address how the data warehouse is built like Bill Inmon did; rather he focused on the functionality of a data warehouse.
Now I shall try to let you understand the definition of Bill Inmon.
- Subject Oriented
The goal of data in a data warehouse is to analyze particular subject area to improve decision making and planning. Such as customers, products, sales, region could be the subject areas. So this means a data warehouse should have a defined scope and it should only stores data under that scope.
- Integrated
The data is the data warehouse is loaded from multiple sources. Each source stores data in different formats and focus on the different aspects of the same subject. So when data coming into data warehouse should be checked, cleansed and transformed into a unified format and projects a single point of truth to allow easy and fast access.
- Time Variant
Every record that is loaded in data warehouse has some form of time variance associated with it. Its mean data warehouse contains historical data. But in an operational system or we can say the transnational system we may hold the most recent records. This way the size of data warehouse grows in size too.
- Non Volatile
Once data have been stored in the data warehouse, it will not be altered or deleted at all. But the exceptions are always there, when we have false or incorrect data get inserted erroneously or the capacity of the data warehouse exceeded and archiving becomes necessary.
Summary
A data warehouse is a concept of data extraction from multiple operational systems, cleaned it, transformed it, cataloged it and then make it available as historical snapshots for ad-hoc queries and use by managers and other business professionals in shape of reporting. This way data warehouse acts as a central repository of the organizational data and ultimately providing a comprehensive homogenized view of the organization. At the end helping forecasting and decision making across the enterprise.
