How to Model Highly Variable Data in a Complex Object Data Model.
Tsin-Shu Yeh, Béatrix de Cambray · 1994
Highly variable data represent an important category of scientific data. The management of such data raises new database problems. These problems are data volumes as well as the great number of states that database operators have to manipulate. In this paper we propose a temporal data model for modeling the variability of numerical data. To this end, a new approach is proposed. We extend the complex object data model to manage highly variable data. To optimize data volume, the approach consists in integrating a description of data variability in complex objects. To manipulate data, this descriptive representation is converted to an internal polynomial representation. Then, queries on highly variable data applies well-known techniques of numerical algorithm techniques on polynomial representations. We present a data representation model and a data manipulation algebra that manipulates temporal states of data and integrates numerical algorithms.