Evolution and Properties of Relational Database Models

Posted on Feb 18, 2025 in Computers

Evolution of Relational Database Systems (1980-1990)

  • Development of large relational systems – SQL commercial.
  • Development of application generators.
  • Distributed systems.
  • Client/server structure (visual languages).
  • Object-oriented relational model.

New Data Representation Models (1980-1990)

  • Problems in the relational model: The semantics of complex items are poorly reflected.
  • Connection of existing data with information in the DB.
  • “Intelligent” unified representation of rules and data, knowledge bases, intelligent information.
  • Existence of databases with unstructured information = BD documentaries information retrieval systems.

New Challenges (1990-2000)

  • New applications:
  1. Treatment of large volumes of image data.
  2. BD for Assistance Systems Design.
  3. The collection of information produced.
  4. BD that support complex information.

New Challenges:

  1. Multimedia technologies BD.
  2. Heterogeneous problems.
  3. Management consultation imprecisely expressed.
  4. Start new ways of accessing the information (data mining).

Relational Model

Created by Codd:

  • Represents data in tables or relations, based on mathematical foundations which gives strength to the operations performed on relations.

Components:

  • Data Structure: Data are organized as relations.
  • Data Manipulation: Powerful language such as SQL is used to manipulate the data stored in the relationship.
  • Data Integrity: Ease to maintain the integrity of the data when these without manipulation.

Properties of Relations

  • The intersection of each row and column is atomic; this implies that there can be no repeating groups.
  • Columns must have the same domain.
  • Each row is unique, guaranteed by the existence of the primary key, which is unique for all occurrences.
  • Each column has a different name, and the column order is unimportant.
  • The order of rows is irrelevant.
  • Relationships must be well-structured, meaning they contain minimal redundancy and can perform insertion, modification, and deletion without generating errors or inconsistency, also called “anomalies”.

Relational Model Details

  • A tuple is defined as the set of values that make up a row in a relationship.
  • A tuple is equivalent to the occurrence of a record within a file. An n-tuple is a tuple composed of n domains, where n is the degree of the relationship.
  • A domain is the set of possible values for a column.
  • Cardinality is defined as the number of tuples of a relation.
  • A relational database is formed by a set of relations or tables, which are associated with each other through the columns that they have in common.
  • Each tuple is composed of one or more domains.
  • It is important to distinguish a domain of an attribute. A column or an attribute represents the use of a domain within a relationship.
  • Each tuple in a relationship is unique and should be possible to define a primary key or identifier that would ensure this uniqueness.
  • The process of manipulating a relational database is far simpler than in the 2 previous models. To represent data in a single form requires only one operator for basic function.
  • rel / arctra = tuple-table-file-field domain registration