SAP: Improving Continuous Top-K Queries over Streaming Data


Consistent top k question over gushing information is a basic issue in the database. In this paper, we center around the sliding window situation, where a ceaseless top k inquiry restores the top k protests inside each question window on the information stream. Existing calculations bolster this sort of inquiries through incrementally keeping up a subset of articles in the window and endeavor to recover the appropriate response from this subset however much as could reasonably be expected at whatever point the window slides. In any case, since all the current calculations are touchy to inquiry parameters and information appropriation, they all experience the ill effects of costly incremental upkeep cost.

In this project, we propose a self-versatile parcel structure to help persistent top k question. It segments the window into sub-windows and just keeps up few hopefuls with most elevated scores in each sub-window. In light of this system, we have built up a few parcel calculations to cook for various protest disseminations and question parameters. To our best learning, it is the principal calculation that accomplishes logarithmic unpredictability w.r.t. k for incrementally keeping up the applicant set even in the worst-case situations.

Leave a Reply

Your email address will not be published. Required fields are marked *