A Fresh Look at Efficient Perl Sorting

Uri Guttman, John Porter, and Larry Rosler

Abstract

Sorting is one of the primary needs of programming, because most
programs produce reports of sorted data.  Efficient sorting therefore
receives a great deal of attention in the Perl documentation, FAQ,
FMTYEWTK, books, and newsgroups.  In this paper, we review and compare
several methods for Perl sorting:  external ('system'),
repeatedly-computed-keys ('naive'), cached-keys ('Orcish Maneuver'), and
mapped-keys ('Schwartz Transform').  We then present a little-known
approach ('single-string key-and-value'), which relies for efficiency on
the default sort algorithm.  This method typically provides much better
performance than the other approaches, and is easy to implement directly
or using a module.