Data that the company owns is one of the most valuable assets available. Specifically, these data and the ability to generate useful business information for decision making is, in the end, the true value, without forgetting that these data sometimes allow the business to be productive, that operations are made in short terms of time and that there are many situations, without which, the company could hardly function efficiently as, for example, all the information based on structured data that allows the management of the supply chain, from products, stocks, orders’ preparation, shipping…
Data are called structured because they are stored logically, with an underlying semantics, usually in tables or tuples of interrelated elements, with restrictions of columns and values (also called fields), which describe not only the information that is derived of them but, on occasion, the restrictions or form of the processes that use them.
The databases that store structured data, in general, are based on SQL technology and the so-called 3 normal forms of databases, also known as normalization. These standards include good practices in the design of structured data or structured databases.
In particular, there are 3 normal forms recognized in the theory of databases that must be fulfilled, to a greater or lesser extent, in professional structured data systems:
- First normal form. This normal form requires that a separate table be created for each set of related data, identifying each set of data related to a primary key. It is recommended not to use several fields in a single table to store similar data.
- Second normal form. It is a next step in the formalization of structured data. In this case, it is recommended to create independent tables for sets of values that are applied to several registers, relating them to foreign key. Records should only depend on a primary key or a composite key if necessary.
- Third normal form. In this third standard, good practice is recommended in structured data to eliminate fields that do not depend on the key, that is, provided that the content of a field can be applied to more than a single record of the table, it is recommended to place these fields in a separate table.
These standards set good practices in the design of structured data and its consequent technological supports such as modern databases or DBMS (Database Management Systems). It is not always possible to have databases of these characteristics, both because there are databases that are imposed by the acquisition of third-party software that manage their own structured data and, sometimes, for convenience or because we focus the structured data on other uses. (such as a Datawharehouse) and these rules are not followed. However, structured data systems that do not comply with these standards can hide bad data design practices and, therefore, jeopardize the success and proper functioning of the applications that use them.
There are other normal forms but, in general, they are not usually considered in real designs of structured data.