| 摘要 |
Processing of data in a higher language database management system is speed enhanced by the use of an external utility or subprogram which determines the amount of free memory available in working memory and selects a plurality of consecutive records from mass memory files which are then transferred as a file unit into a virtual record storage buffer in the working memory leaving just sufficient free working memory to accomplish required processing. The data in the records in the virtual record storage buffer is then continuously processed in the normal manner but without any time consuming intermediate selection of records. In improved versions of the method, the records are read initially into a primary buffer and process records are output through a secondary buffer which is desirably the same as the primary buffer. The utility is written in assembly language for additional speed.
|
