Space efficient data structures for dynamic orthogonal range counting

  • Authors:
  • Meng He;J. Ian Munro

  • Affiliations:
  • Faculty of Computer Science, Dalhousie University, Canada;Cheriton School of Computer Science, University of Waterloo, Canada

  • Venue:
  • Computational Geometry: Theory and Applications
  • Year:
  • 2014

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a linear-space data structure that maintains a dynamic set of n points with coordinates of real numbers in the plane to support orthogonal range counting in O((lgnlglgn)^2) worst-case time, and insertions and deletions in O((lgnlglgn)^2) amortized time. This provides faster support for updates than previous results with the same bounds on space cost and query time. We also consider the same problem for points on a UxU grid, and design the first succinct data structure for a dynamic integer sequence, S, to support range counting queries defined as follows: Given a range, [i"1..i"2], of indices and a range, [v"1..v"2], of values, count the number of entries of S[i"1..i"2] that store integers from [v"1..v"2].