Block concatenated code word surrogate file for partial match retrieval

Abstract

In this paper, a block concatenated code word (BCCW) surrogate file scheme is developed to speed up partial match retrieval operations. A BCCW is generated for each block of the data file by hashing the attribute values in the data block. Then the BCCWs forms a surrogate file which is used as an index to the data file. For a partial match retrieval query, a block query code word (BQCW) is generated and compared with the BCCWs. Only those data blocks whose corresponding BCCWs match the BQCW are retrieved from secondary storage and compared with the actual query. The size of the BCCWs is usually less than 10% of the size of the data file and only a subset of each BCCW is accessed. Thus, we can obtain considerable speed up in partial match retrieval by using the BCCW surrogate file. The storage requirement and the performance of the BCCW surrogate file are evaluated and compared with those of other schemes.