Multidimensional databases are generally used to create OLAP applications. They are made up of several tables of facts and dimensions. In this way, each dimension table contains a simple primary key that in turn composes the primary key of the fact table. In addition, multidimensional databases allow quick access to data that has already been added, according to the user’s needs. However, it is worth developing the modus operandi of this type of database.
Multidimensional data bases
Generally, they use OPAP hypercubes, and the data is represented in these hypercubes to “n” dimensions. That is, the data is classified according to different structures or analyzes, such as tempo, space, etc. The main advantage of using OLAP hypercubes is that it allows the level of abstraction of the data to be increased. There are different types of multidimensional bases and then we will analyze the main most used models: MOLAP, ROLAP, and HOLAP.
MOLAP (Multidimensional Online Analytical Processing), stores the data in a cube.
The objective of this database is to precalculate all possible crossings in order to restore the data as quickly as possible. Thus, MOLAP, presents certain advantages, but also some disadvantages that must be considered. As for the advantages, we can mention the gain of time that is obtained in the restitution of data, since this is done instantaneously. However, as regards the drawbacks, it must be considered that MOLAP is quite expensive, since licenses are required for its use, and the development of cubes also requires high economic costs.
ROLAP (Relational Online Analytical Processing), obtains the data through relational tables, which allows complex and extensive requests to be resolved. However, the results are not stored, so whenever the user needs to consult the results, he must re-launch the request. Unlike with MOLAP, as the main advantage, ROLAP presents a low cost since it uses already existing resources. However, contrary to MOLAP again, ROLAP has a longer response waiting time.
HOLAP (Hybrid Online Analytical Processing) uses the two previously explained systems, MOLAP and ROLAP. That is, HOLAP allows the user to access the data report of a cube, but also has access to a second more detailed report that contains the data from the tables. The main advantages of the HOLAP multidimensional database are, first of all, that it requires less investment and therefore has a lower cost than the MOLAP base. Also, the response time is fast. However, HOLAP is not full of drawbacks, and among them, it is worth mentioning the impossibility of using this database in the event that the data present a high degree of complexity, or that they present a high number of cross-referenced data.
In conclusion, the main purpose of the multidimensional databases is to regroup, organize and order information or data, which come from different sources, in order to store them and perform an analysis of the information as quickly as possible. In addition, there are different types of multidimensional databases, and each has different advantages and disadvantages that must be assessed when making the relevant choice according to the needs of each project.