Top-k Query Processing with Conditional Skips.

This work improves the efficiency of dynamic pruning algorithms by introducing a new posting iterator that can skip large parts of the matching documents during top-k query processing. Namely, the conditional-skip iterator jumps to a target document while skipping all matching documents preceding the target that cannot belong to the final result list. We experiment with two implementations of the new iterator, and show that integrating it into representative dynamic pruning algorithms such as MaxScore, WAND, and Block Max WAND (BMW), reduces the document scoring overhead, and eventually the query latency.