Data Abstraction Technique for Relational Database Query Processing

Abstract

AbstractThis paper proposes a schema conversion approach for relational database query processing. The approach not only maintains high data independence by encapsulating the data and its physical structures for a query with the procedures that manipulate them, but also aims to achieve high‐speed processing by dedicating the procedures to individual functions. This approach is experimentally compared with the interpretive and compilation methods in terms of data independence, performance and the procedure program size. The following conclusions are drawn from the experiments.(1) Although the effect of the change in physical structure is greater than that of the interpretive approach in regard to data independence, only a limited number of parts are affected and they are easily reorganized by each dedicated procedure. (2) The performance is extremely close to that of the compilation approach regardless of the kind of computer and programming language. Furthermore, the overhead for calling the procedures can be further reduced with the use of microprograms and hardware. (3) Although the number of specialized procedures increases, this is compensated by the relatively small program size of each procedure.