Event-Oriented Data Models and Temporal Queries in Transaction-Time Databases

Abstract

Past research on temporal databases has primarily focused on state-based representations and on relational query language extensions for such representations. This led to many different proposals that had did not succeed in making a significant impact on SQL-compliant DBMS.More recently however, there has been significant interest and progress on event sequences,leading to vendor-proposed extensions of SQL standards for pattern queries based on Kleene-closure expressions.In this paper, we first outline these extensions and their uses in dealing with sequence of events, and then show that they can also be used effectively to express more traditional temporal queries, such as coalescing and joins, on state-based representations.Thus, we propose an approach that takes full advantage of the fact that every state-based representation also has a dual representation based on its start-event and its end-event.