摘要 |
<p>A database comprises a sequential file 15 and a transposed file 11 in external memory. A query specifies various keywords, which are used to access the transposed file. Data concerning the various keywords, and especially identifiers for retrieval objects containing the keywords, are loaded into main storage. Inside main storage this information is rearranged, so as to be able to rank the retrieval objects in terms the keywords they contain. Identifiers from this ordered list can then be used to access information from the sequential file, with only retrieval objects with high query scores being accessed.</p> |