Independent Data Marts
An independent data mart has neither a relationship with the enterprise data warehouse nor with any other data mart. Its data is input separately and its analyses are conducted autonomously. Because the data is not derived from the central warehouse, the likelihood that it does not match the enterprise data is high. Which version of reality is correct? How can a user know?
Teradata often discourages the use of independent data marts, sometimes referred to disparagingly as “data basements.” Implementation of independent data marts is antithetical to the motivation for building a data warehouse in the first place: to have a consistent, centralized store of enterprise data that can be analyzed in a multiplicity of ways by multiple users with different interests seeking widely varying information.
A data basement is a collection of independent data marts. Suppose you have parts that you decide to store in your basement. There is no particular rhyme or reason to what part is stored or where it is stored other than convenience. Continuing the analogy, what is stored in the basement depends on what any family member decides needs to be stored there. If you need to locate a part that you think might have been stored in the basement, you ask everybody in the family if they have seen it recently and then you make your search based on their recollections. If you need to visit more than one basement to find your parts, it is unlikely they will be compatible even if you are able to find them.
This method of storing data is essentially the same as the mix of paper databases and mixed hierarchical and relational online databases spread among multiple departments that supports many businesses today. It is the sort of situation that businesses generally want to escape, not automate.