Large companies often find that subdividing the data warehouse into smaller, more manageable components gives them more flexibility in the use of the data. To this end, smaller subsets of data are set aside called data marts. Data marts can be created to warehouse data from just one department within the company. Information can also be grouped from several related departments to form a data mart.
A data mart should have all the same “rules” as the overall data warehouse, but with smaller parameters of included information. In this instance, only those records relating to the data mart subset would be included, and all other data would be excluded. But don’t be misled. The excluded records can still be held in a separate data mart and still function as part of the entire data warehouse system. Smaller subsets, or data marts, should make information more manageable, not less. With proper organization and maintenance, data marts can make unwieldy information more accessible, and thus, more powerful.
Think of it this way. When you move into a new home, you organize your belongings in a logical manner. You do not store dinner plates in the bedroom because accessing them would be bothersome. Instead, you keep dinner plates and all other kitchen items in the kitchen. You may even decide to further organize your kitchen into two realms by storing your dinnerware, place settings, and serving pieces in a china hutch in the dining room and only store cooking and preparation tools in the kitchen. These divisions make the use of the kitchen more practical and efficient. The use of data marts works in the same manner. By compartmentalizing data, that data becomes more efficient and accessible.