Academic journal article Informatica Economica

Contributions to Logical Database Design

Academic journal article Informatica Economica

Contributions to Logical Database Design

Article excerpt

This paper treats the problems arising at the stage of logical database design. It comprises a synthesis of the most common inference models of functional dependencies, deals with the problems of building covers for sets of functional dependencies, makes a synthesizes of normal forms, presents trends regarding normalization algorithms and provides a temporal complexity of those. In addition, it presents a summary of the most known keys' search algorithms, deals with issues of analysis and testing of relational schemes. It also summarizes and compares the different features of recognition of acyclic database schemas.

Keywords: Logical Database Design, Functional Dependencies, Normal Forms, Acyclic Database Schema

(ProQuest: ... denotes formulae omitted.)

1 Introduction

A properly designed database provides access to accurate and updated data. Because a correct design is essential for achieving the goals of using a database, the ability to design databases and associated applications is critical to the success of any modern enterprise. Automatic drafting and the development of database and of the information system as a whole consists of a series of steps that require methodological research, efficient models and algorithms and software design tools.

Within the logical design phase, the conceptual schema, expressed in a high-level data model is transformed into a global logical schema, described in a logical data model, for example, the relational model, without taking into account a specific DBMS, in this case, is obtained a system independent logical design, but dependent on the data model.

The global logical schema is normalized [16], [17], all keys [40] and the links between relations are identified. Then, the global logical schema and the information about access to data serve as input for the next step which is distribution design [47]. The objective of this phase is to design local logical schemes, which are distributed to all the stations of the distributed system. The paper reviews the current issues that pertain to logical database design for the purpose of automation of this process.

2 Visions on functional dependencies inference models

It is being considered [52] that the problem of logical design of relational database consists in laying the theoretical and practical basis for taking decisions about:

* What should the relations of the database;

* What attributes should each relation consist of.

Here a design aspect needs to be mentioned - determination of integrity constraints [49]. Constraints, however, are withdrawn at the conceptual design phase. Therefore, the success of DBMS utilization with up-to-date mechanisms to maintain integrity constraints cannot depend on a common treatment process for obtaining these constraints [58].

The simplest integrity constraint is functional dependency and Codd introduced it into the database theory [15]. Deriving from semantic rules, which translate the restrictions of the domain of interest, the designer has to define, among other types of dependencies, and functional dependencies and to introduce them in the definition of the database schema. The properties of dependencies are the ownership of the relational scheme (intensions) of the database, and not of any extension of the database, that is these dependencies are invariant and need to be satisfied by all legal extensions that correspond to the scheme [18]. The only way to find the valid functional dependencies for a schema consists in a careful analysis of each attribute's significance and of the way values are assigned to the attributes [51].

Starting from a set of functional dependencies, attached to a relational scheme, other valid functional dependencies can be deducted. There are many rules of inference and in order to be able to make a formal presentation of these, three of them were chosen by Armstrong [1], and the rest of them are derived from them [35]. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.