Identifier
“An Identifier is a set of one or more Attributes that uniquely distinguishes each instance of a Class.” [MB]
We can clarify the concept of uniqueness further. Given a populated Class and a value for each Attribute of some Identifier of that Class, at most one Instance of that Class can be selected using that set of Identifier Attribute values.
An Identifier is not simply a means of locating an instance. Identity is a constraint that ensures that no two instances share certain values. To express a rule that two chess pieces may not be placed on the same square, for example, tag the attribute pair (rank, file) as an Identifier.
- Number + Class + Domain
Identifiers are numbered uniquely within a Class.
A Class’s Identifiers are numbered sequentially starting from one. By convention, the first Identifier, I1 is shown as {I}. Subsequent Identifiers are {I2}, {I3} and so on. There is nothing special about I1 or what is informally referred to as the primary identifier. Though it is typically the first or default choice to use in Attribute References to the Identifier's Class, any other Identifier of the same Class will do. When a modeler chooses to enforce a constraint by merging a referential attribute, an Identifier other than the primary might be chosen if it houses a corresponding attribute.
Type: ID Number I<n> I1, I2, ... , based on Nominal