Business Intelligence: the concept of ‘Relationship’

There are several ways to organize data in a database. A particular method, useful for structuring information in a DB, is called Logical Data Model (LDM). A Logical Data Model specifies the rules according to which data are structured (DDL = Data Description Language) and the associated operations that are allowed (DML = Data Manipulation Language). rnTo download the full text and for more information, click HERE

Ultimately, a Logical Data Model consists of a DDL plus a DML. The LDM can also be seen as a technique for the formal description of the data structure, usage constraints, and operations. The tools available vary from one Logical Data Model to another. We can say that each DBMS is associated with a particular LDM. The LDM corresponds to the logical level of data representation, while the hardware and software environment correspond to the physical level of data representation. In fact, in database theory, there are generally three distinct levels referenced: • conceptual level or schema • logical level or schema • physical level or schema The conceptual schema is independent and completely detached from the computer; it can be developed without any particular knowledge of computer science, but only with a model of structuring reality in mind. The logical schema of the database, instead, takes into account the adopted model (e.g., relational model). Finally, the physical schema concerns the actual realization of the computer system and is therefore closely connected to the type of computer. More formally, an LDM is a combination of at least three components: • a set of data structure types • a set of operators or functions that can be applied to any valid instance of the defined data types • a set of general integrity rules (which implicitly or explicitly define the set of states consistent with the DB, the changes of states, or both) There can be two or more different DBMSs supporting the same Logical Data Model, so knowledge of at least one Logical Data Model is sufficient to develop database applications. Care must be taken not to confuse the two terms “information model” and “data model”: an information model is a description of the real world in terms of a Logical Data Model. A simple approach is given by the relational data model. The Relational Model was proposed by Dr. Edgar Codd in 1970. The relational database model specifies a data structure through so-called “Relations” and some high-level languages for manipulating relations. The “relation” is a mathematical term identifying a two-dimensional table; in other words, according to the relational model, information is organized in rows and columns. The reason for calling such a table a “relation” rather than a “matrix” derives from the fact that data contained in the same columns are homogeneous, whereas data present in the same rows are not. Each category (or entity) is described by a particular relation, so each relation has a unique name. A set of related facts are represented in a row of the relation, which is called a “tuple.” In other words, each tuple of a relation corresponds to an instance of a particular entity. All entries present in a column of a relation are of the same type. Each column is assigned a different name called an attribute, as already noted concerning the Entity-Association model. The term “Domain” refers to the set of possible values that an attribute can assume and which we find (all or in part) as values in the field corresponding to that attribute. Distinct attributes can have the same domain. ALL RIGHTS RESERVED

Pubblicato in Business