A multidimensional database is a form of database that is structured to optimize online analytical processing (OLAP) and data warehouse applications. It can receive data from a variety of relational databases and structure the information into categories and sections that can be accessed in a number of different ways. Even persons who have relatively little experience working with a database often find that a multidimensional database, or MDB, requires only a short time to master.
While just about every relational database is structured for keyword searches and building a query by specifying fields and perimeters, the multidimensional database goes one step further. Rather then building a query, a user simply poses the question in everyday verbiage. This approach is used with several online help tools associated with software programs such as word processing and spreadsheet applications, as well as several of the more popular search engines currently in use.
When it comes to using a multidimensional database for internal business purposes, the main advantage is the ease of obtaining data quickly and succinctly. For instance, if an end user wanted to determine how many widget sales were generated during the third quarter of the past year in a given sales territory, this data could be obtained from a multidimensional database with a simple question. By asking “How may widgets were sold during third quarter 2007 within the Southwest Territory?” the end user does not have to go through the steps of building a report, specifying fields, and restricting content within those fields to selected criteria. The use of one simple question will accomplish the task.
The exact means of formulating a question will determine on several factors. One key element is the sorting and type of data within a multidimensional database. Any question that contains a request for information that is not found within the database will not result in a direct response, although many MDBs will respond with data based on what data that is available. This means that if an end user asks “How many widgets were sold in Virginia last month?” the database will only be able to provide a focused response if each sale includes the state of origin and the database was configured to understand which month and year “last month” indicates. If not, the multidimensional database will probably respond with some partial answers, or ask for clarification.