Research On Algorithms About Temporal Aggregation Query

Temporal database systems extend traditional relational database systems. One important such extension is to support time-varying aggregation. For its increasing importance in the data warehouse applications, temporal aggregation query have become the focus of the temporal database research.In this thesis, we first introduce the consensus glossary and data models of temporal database systems. Then we give some introduction of the taxonomy of aggregation, the support for temporal aggregation in temporal query language. Later we explore the newly algorithms of temporal aggregation, which are aggregation tree algorithm, k-ordered aggregation tree algorithm, 2-3 tree algorithm, balanced tree algorithm and merge-sort aggregation algorithm. On the basis of analysis and comparison of these algorithms we make some improvement to these algorithms. Through storing time point value instead of time period information in the leaf of the aggregation tree, we efficiently decrease the memory requirement of aggregation tree algorithm; Through paging out part of the aggregation tree, the paging aggregation tree algorithm we provide can make the aggregation tree algorithm suit for the situation of limited memory; Through providing a new way to compute the aggregation values of balanced tree algorithm, we make the balanced tree algorithm suit for the situation when the valid time periods of tuples are closed-closed timestamps. Analysis and experiments show that our improvement can efficiently improve the performance of original algorithms and greatly extend the situation they are applied to.