A temporally oriented data model

Abstract

The research into time and data models has so far focused on the identification of extensions to the classical relational model that would provide it with “adequate” semantic capacity to deal with time. The temporally oriented data model (TODM) presented in this paper is a result of a different approach, namely, it directly operationalizes the pervasive three-dimensional metaphor for time. One of the main results is thus the development of the notion of the data cube: a three-dimensional and inherently temporal data construct where time, objects, and attributes are the primary dimensions of stored data. TODM's cube adds historical depth to the tabular notions of data and provides a framework for storing and retrieving data within their temporal context. The basic operations in the model allow the formation of new cubic views from existing ones, or viewing data as one moves up and down in time within cubes. This paper introduces TODM, a consistent set of temporally oriented data constructs, operations, and constraints, and then presents TOSQL, a corresponding end-user's SQL-like query syntax. The model is a restricted but consistent superset of the relational model, and the query syntax incorporates temporal notions in a manner that likewise avoids penalizing users who are interested solely in the current view of data (rather than in a temporal perspective). The naturalness of the spatial reference to time and the added semantic capacity of TODM come with a price—the definitions of the cubic constructs and basic operations are relatively cumbersome. As rudimentary as it is, TODM nonetheless provides a comprehensive basis for formulating an external data model for a temporally oriented database.