On Thu, 17 Jun 1999, Richard K. Moore wrote: > 6/16/99, Ronald J Kimball wrote: > >If you want to sort the entire file by line, you have to read the entire > >file into memory. You can then use a Schwartzian Transform to efficiently > >sort the lines: > > I'm curious as to why people haven't been advising the use of a hash (or > two). Alan's application seems to be one with only a small amount of data. > Is the Schwartzian Transform incredibly faster as a solution for large > amounts of data? > It's a pretty standard way to go these days, and it won't let you down. If you're willing to tweak you can usually come up with somewhat faster sorts for particular problems. I benchmarked the provided S.T. over 1000 iterations with the problem data vs. 1000 iterations of a sorting solution that used sprintf() to line the first name and surname up in a single string, and sorted _that_, and the latter was about 25% faster. But I lost the speed gain in recreating the original format for printout, so... Normally I'd just use the S.T. myself. Arved ===== Want to unsubscribe from this list? ===== Send mail with body "unsubscribe" to macperl-request@macperl.org