A query processing method for integrated access to multiple databases

Abstract

Multiple autonomous databases within a network system often contain similar information although their schemata are heterogeneous. This paper considers such heterogeneity and proposes a new method for achieving integrated access to multiple databases having similar schema. The method is based on the construction of a “virtual database (VDB)” by making use of null values and “virtual functional dependency (VFD),” which are formed over physically distributed component databases and applied to the query processing. We could obtain additional information relevant to the query and proved that this information is logically valid. When the VFD formed over a VDB is acyclic among the nodes that the query refers to, sound and complete answers can be efficiently calculated. We developed and implemented a bibliographic information retrieval system employing the schema integration method described herein such that integrated access is obtained for bibliographic information written in BibT_EX.