| 摘要 |
<p>A data sorting method and system based on a MapReduce model. The sorting method comprising the following steps: when each MAP module sorts elements in a data block in a MAP phase, each time a sorting result of one element is determined, adding the element to a first-in-first-out data structure; in a Reduce phase, a Reduce module acquiring all current elements in the first-in-first-out data structure; acquiring a current global extremum element; outputting the current global extremum element as a sorting result; removing the output current global extremum element from a Reduce-phase element set, and updating the Reduce-phase element set. With the method and the system, without the need of waiting until a sorting process in a MAP phase ends, an extremum element can be selected from current elements in a first-in-first-out cache queue in a Reduce phase, and a sorting result is output in a streamed manner. In this way, the sorting result can be output in advance, so as to reduce the waiting time of the user.</p> |
