Summary: PREPRINT. In Proc. 2nd Annual European Symposium
onAlgorithms, pages 8293. Springer Verlag, 1994.
Faster Searching in Tries and Quadtrees---
An Analysis of Level Compression \Lambda
Arne Andersson Stefan Nilsson
Department of Computer Science, Lund University,
Box 118, S221 00 Lund, Sweden
We analyze the behavior of the levelcompressed trie, LCtrie, a compact
version of the standard trie data structure. Based on this analysis, we argue
that level compression improves the performance of both tries and quadtrees
considerably in many practical situations. In particular, we show that LCtries
can be of great use for string searching in compressed text.
Both tries and quadtrees are extensively used and much effort has been
spent obtaining detailed analyses. Since the LCtrie performs significantly
better than standard tries, for a large class of common distributions, while
still being easy to implement, we believe that the LCtrie is a strong candidate
for inclusion in the standard repertoire of basic data structures.
A fundamental, well known, and very well studied technique for storing and retriev