Compressed Indexes for Approximate Library Management

Abstract

This paper investigates the approximate library management problem, which is to construct an index for a dynamic text collection $L$ such that for any query pattern $P$ and any integer $k$, we can report all $k$-error matches of $P$ in $L$ efficiently. Existing work either focussed on the static version of the problem or assumed k=0. We observe that by combining several recent techniques, we can achieve the first compressed indexes that support efficient pattern queries and updating simultaneously.