COSI 127b Lecture Notes - Lecture 4: Data Integrity, Functional Dependency, Unique Key
8 Mar 2020
School
Department
Course
Professor
Document Summary
Instead of superkeys, weak entity sets use discriminators. Aggregation makes an entity set out of a relationship set. Attributes can be simple, multivalued, composite, or derived. The primary key is automatically determined for relationship sets. In user defined subclasses, membership is explicitly determined. In condition defined subclasses, a predicate is associated with the subclasses. An overlapping subclass can belong to one or more entities, whereas in a disjoint subclass they belong to exactly one entity set. The first step of the canonical cover algorithm is to apply the union rule. Normalization starts with all attributes in one table, the universal relation. Next, determine the functional dependencies by using the canonical cover algorithm. The three goals of decomposition are: lossless joins, redundancy avoidance, and dependency preservation. The goal of lossless joins seeks to avoid information loss, and is the most important goal of the three goals of decomposition.
