Some Issues of Normalization of Relational Database

Abstract

Building a logical model of a relational database involves: defining real or imaginary entities and their attributes,
revealing functional dependencies between attributes, creating relationships among the relations, detecting undesirable
anomalies during data insertion, modification and deletion. The process of normalization is used to bring relations to certain
canonical (normal) forms. Each higher normal form satisfies the properties of the lower normal form and also has better
properties than it. Normalization ensures the reduction of data redundancy, the removing of undesirable functional
dependencies and the integrity of data. There are considered: First, Second, Third, Boyce–Codd, Fourth, Fifth, Sixth normal
forms. The normalization process is optional; a database specialist intuitively constructs already normalized relations, but
discussing the process allows us to understand what problems we will encounter when working with the database. The article
discusses examples of normal form of relations. Key attributes of relations and various functional dependencies between
attributes are defined. The projection and natural join operations of relational algebra are discussed for relational database
normalization. The issues discussed in the article have theoretical, educational-methodological and practical significance. The
theoretical research, data comparison and analysis, induction, deduction methods are used.