A data store is very a very important aspect of a data warehouse in that it acts as support of the companies need for up-to-the-second, operational, integrated, collective information. It is a place where data such as databases and flat files are saved and stored. Data stores are great feeders of data to the data warehouse.
In a broad sense, a data store is a place where data is integrated from a variety of different sources in order to facilitate operations, analysis and reporting. It is can be considered an intermediate data warehouse for databases despite the fact that a data store also includes flat files.
Some data warehouses are designed to have data loaded from a data store which consists of tables from a number of databases which supported administrative functions like financial, human resource, etc).
In some cases, the store are contained in one single database, while in other cases, the data store is scattered in different databases in order to allow tuning to support many different roles.
Those who prefer not to store a data store in a single database argue that the tuning choices are based on the very nature of the data and not on database design and the access on the large volumes of data would be negatively affected to a certain degree. It also matters in terms of the politics of getting everyone's concurrence.
A data store is an important link in a data warehouse's staging area. A staging area is conceptual place in the data warehouse which stands between the analytics system and the legacy system.
Some people think of the staging area as the "back room" portion of the data warehousing environment. This is where the collective process of the data warehouse known as ETL (extract, transform and load) is done. Whenever there is a need for data to be executed in the ETL process, the data warehouse gets the data from the data store which contains all the data at rest.
The data store, being an integral part of the data warehouse architecture, is the first stop for the data on its way to the warehouse. The data store is the place where data is collected and integrated and made sure of its completeness and accuracy.
In a lot of data warehousing implementations, all data transformations cannot be completed without having a full set of data being available. So, if there is a high rate of data, capturing is possible without having to constantly change data in the warehouse.
In general, data stores are normalized structures which integrate data which are based on certain subject areas and not on specific applications. For instance, a business organization may have more than 50 premium applications.
A premium subject area data store collects data using a feed from different applications providing near real time enterprise wide data. The data store is constantly refreshed in order to stay current. The history will then be sent to the data warehouse.
A data store can be a great tool as a reporting database for line-of-business managers and service representatives who will be requiring an integrated picture of the enterprise operational data. Some important aspects of business operation such as operational level reports and queries on small amounts of data can be made more efficient by the data from the data store.
For instance, if one wants specific data for only one calendar quarter, it may be wise to just query the data store. It would be much faster because querying the data warehouse will involve sifting through data for several years.