Redundancy in Databases

Databases

What is redundancy? Why is it a problem? How can it be reduced?

A systematic collection of related data that facilitates data management

Designed, built, and populated with one or more specific purposes in mind

Key problem with storing data multiple times in a database:

Redundant data

Updates have to be performed multiple times = effort is multiplied

Storage space is wasted which may be problematic for large databases

Files that represent the same data may become inconsistent

Increases possibility of errors and continuous errors

Can slow down performances especially with complex queries

Problem caused by many-to-many relationships

e.g.

In database with student records: every student can belong to multiple departments, every department can have multiple students

Can be solved by creating separate tables

Common method to reduce redundancy is to normalise database tables

Ensures each data cell is unique

Relational tables = so they can be linked

If data is stored in different places it's harder to update

BUT

Can be thought of as a policy describing a multi-step process towards normalising data and solving redundancy

Table with fields containing more than one kind of information = fields are 'non-atomic' i.e. problematic

Searches for particular elements have to take into account the multiple entires

Searching for 'French History' within a field might bring results of French, History, and French History

Changes have to be done in many places which would be difficult to find

e.g.

Need to be brought to an atomic level

therefore

Rules are divided into: First Normal Form / Second Normal Form / Third Normal Form / Boyce-Codd Normal Form

A table is in the First Normal Form if all underlying fields are atomic and there are no repeating groups and fields

Normalisation

E.g. changing a student's address when student appears more than once

Repeated entries in tables are a waste of time, effort, space

Each set of columns must have a unique value

Each row should have a primary key that distinguishes it as unique

= each record is unique