Constrained Relation Browse

Introduction

The purpose of constrained relation browse is to facilitate the building of queries which, upon execution, do not produce empty result relations. The relational algebra of CRB is the method of determining what values are allowable for a given query attribute. Thus, as this is a demonstration of CRB, it is never possible to select a query in the interface below which would cause an empty result relation.

A constrained relation query begins with an empty query. In this example, each attribute "drop-down" of the query relation is populated with all possible values. A change in one query attribute results in a change in the available values for all other query attributes. Attribute "drop-downs" are only populated with values which, when selected, would not produce empty result relations.

Learn about Relational Algebra

Wikipedia has an excellent introduction to relational algebra. To understand this article, you need only know about the "project" and "natural join" operators.

Formulas

Let Q₁ be a relation of no more than 1 tuple with no null attribute values. Let a₂ be the set of unique attributes in R₁ which are not in Q₁. Thus the unique relation for any attribute a in a₂:

π_a(R₁ ⋈ Q₁)

Let a₁ be the set of common attributes in relations R₁ and Q₁. For any attribute a in a₁ there is another set a₃ that is the set a₁ excluding a. Thus the unique relation for a in a₁:

π_a(R₁ ⋈ π_a3(Q₁))