| Abstract
|
Research on temporal relational databases … Research on temporal relational databases can be classified into two categories by the type of time-stamping method used: tuple time stamping and attribute-value time stamping. Although the former method of time stamping retains the simple structure of the relational model, standard relational operators do not form either a closed or a complete algebra for manipulating such temporal objects. In this paper, we provide a proper algebra for historical relations with interval-stamped tuples. To formalize the representations of and operations on time, we first develop a general model of time that includes a set of operations on chronons and intervals. We then define the temporal semantics of a unisorted algebra that is closed for historical relations. We ensure that our algebraic operators can support historical queries by requiring the algebra to support all operations in the time model. In contrast to time stamping of attribute values, time stamping of tuples uses the classical normalized relation that is supported by most relational database technologies. We discuss briefly how we implement the operations in our algebra in such a database with a temporal query processor for SQL. e with a temporal query processor for SQL.
|
