Key

Synonyms

Uniqueness constraint

Definition

In the relational model, a key for a relational schema is a set of attributes whose value(s) uniquely identify a tuple in a valid instance of the relation. Said another way, key value(s) appear at most once in a relation. Often the set of attributes constituting a key is a set with a single attribute. For example, in the relation

$$ \mathrm{Customer} \frac{\left(\mathrm{Customer} \mathrm{ID} \mathrm{Name} \mathrm{Address}\right)} {\begin{array}{l} 11111 \mathrm{Pat} 12\ \mathrm{Maple}\\ {}22222 \mathrm{Tracy} 44\ \mathrm{Elm} \end{array}} $$

the singleton set consisting of just CustomerID is a key and so is the set consisting of the pair of attributes (Name, Address). The values of CustomerID (here, 11111 and 22222) each uniquely identify a tuple. In any valid instance for this relation, no CustomerID value may appear twice. Similarly, the pair (Name, Address) is a key. Pairs of (Name, Address) values (e.g., <Pat, 12 Maple>) uniquely identify...