Organization of quasi-consecutive retrieval files

摘要

In 1972, Ghosh introduced the consecutive retrieval (CR) file organization. It is an efficient file organization in which all records pertinent to a query are consecutively stored on linear storage locations.In this paper, we introduce the quasi-consecutive retrieval (QCR) file organization. The QCR file organization is an extension of Ghosh's CR file organization and normally offers less redundancy. In the QCR file, all the records pertinent to each query are not necessarily stored consecutively, but rather they are stored within an area of the buffer size.In this paper, we discuss graph theoretic properties of CR files and QCR files by using the properties of interval graphs well known in graph theory. We provide a basic condition for the existence of a QCR file without redundancy for a given buffer size. By introducing redundant queries, such a condition is simplified.Furthermore, a heuristic computer algorithm is given to organize a QCR file with less redundancy for a given buffer size.