Access Methods for Bi-Temporal Databases

Abstract

While much work has recently appeared in literature on access methods for transaction-time databases, not much has been done for indexing bitemporal databases, i. e., databases that incorporate both transaction and valid time dimensions. In this paper we first discuss the issues involved in addressing general bitemporal queries and then propose two general approaches in solving such queries. For simplicity we present our findings in relation to the so-called bitemporal pure-timeslice query. However our methodology applies to more complex bitemporal queries. The first approach reduces bitemporal queries to partial persistence problems for which an efficient method is then designed. Using this approach we introduce a new access method for the bitemporal pure-timeslice query, the Bitemporal Interval Tree. The second approach “sees” bitemporal data objects as consisting of two intervals, a valid-time and a transaction- time interval. It then divides bitemporal data objects in two categories according to whether the right endpoint of the transaction time interval is known, and uses a different, R*-tree based organization, for each category. In this paper we present the advantages and disadvantages of these approaches. In addition we compare them through simulation with a straightforward approach that “sees” the intervals associated with a bitemporal object as one rectangle that is stored in a single R*-tree.