Query Processing Using the Consecutive Retrieval Property

Abstract

In this chapter, we will discuss procedures to handle multiple queries utilizing the consecutive retrieval property. A set of queries is said to have the consecutive retrieval property if the records pertinent to every query in the set are consecutively located in a linear structure file, so that each query can be processed by a single direct access of the file. The consecutive retrieval property can be used to reduce the computation time as well as the storage space. There are many cases when a given query set does not satisfy the property and we need to develop procedures to handle such cases. One approach is to generalize the file structure. Instead of the linear structure various graph structures can be used. Another approach is to permit multiple accesses for queries with low usage frequencies. We give a greedy algorithm to construct a consecutive retrieval file where the preference of queries are given by the frequencies of usage. File organizations with different main memory buffer sizes are also discussed, such as the quasi-consecutive retrieval file organization and the buffer-limited quasi-consecutive retrieval file organization. In these organizations records corresponding to each query are not required to be stored consecutively; instead these records are to be located within the size of the buffer. A multiple query processing method for cost reduction by keeping some records in the buffer is also described.KeywordsBoolean FunctionQuery ProcessingDirect AccessBuffer SizeInterval GraphThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.