Efficient Implementation of the TMDA Operator

Motivation: Most real-world database applications, e.g., in the financial, medical, and scientific domains, manage temporal data, which is data with associated time intervals that capture some temporal aspect of the data, typically when the data were or is true in the modeled reality. In contrast to this, current database management systems offer precious little support for temporal data management.

An important operator for temporal databases is aggregation, which transforms an arguemnt relation into a summary result relation. Traditionally this is done by first partitioning the argument relation into groups of tuples with identical values for one or more attributes, then applying an aggregate function, e.g., sum or average, to each group in turn. For temporal databases, aggregation is more complex because the interval values can also be used for defining the grouping of argument tuples. The Temporal Multi-Dimensional Aggregation (TMDA) operator [1] is an expressive temporal aggregation operator which generalizes a variety of previously proposed aggregation operators.

Problem: An efficient implementation of the TMDA operator poses new challenges and requires sophisticated data structures and algorithms. In particular, the following aspects are crucial for the performance of TMDA:

• implementation of the group table
• implementation of the endpoint tree
• evaluation of the theta-condition

Objective: The aim of this thesis project is to study, develop, implement and evaluate an efficient solution for one (or more) of the above mentioned aspects.

Reference:

1. M. Böhlen, J. Gamper, and C.S. Jensen. Multi-dimensional aggregation for temporal data. In Proc. of the 10th International Conference on Extending Database Technology (EDBT-06), LNCS 3896, pages 257–275, Munich, Germany, March 2006. (Abstract)

Contact person: Johann Gamper (gamper_at_inf.unibz.it)