Abstract |
Sequence representations supporting queries access, select and rank are at the core of many data structures. There is a considerable gap between different upper bounds, and the few lower bounds, known for such representations, and how they interact with the space used. In this article we prove a strong lower bound for rank, which holds for rather permissive assumptions on the space used, and give matching upper bounds that require only a compressed representation of the sequence. Within this compressed space, operations access and select can be solved within almost-constant time. |