http://www.perlmonks.org?node_id=382716


in reply to Finding longest palindrome from a string

I'm insanely proud of this one as it represents the most interesting bit of coding I've had to do in a while. It even seems to work, although I did cheat by importing POSIX.

#! /usr/bin/perl -w use strict; use POSIX qw(ceil); sub elgon { my $string = shift; my @answers; my %char_hash = map { $_ => 1} split //, $string; foreach my $key (keys %char_hash) { my @appearances; for (my $i = 0; $i < length($string); $i++) { push (@appearances, $i) if substr($string, $i, ++ 1) eq $key; } foreach my $start(@appearances) { foreach my $finish(reverse @appearances) { next if $start >= $finish; my $half_length = ceil(($finish - $sta +rt + 1) / 2); push @answers, substr($string , ($star +t) , ($finish - $start + 1) ) if substr($string, $start , $half_length) eq reverse substr ($string, ($finish - $half_length + +1), $half_length); } } } return "FAILED!" if ! scalar(@answers); my $longest = ""; map { $longest = $_ if length($longest) < length($_) } @answer +s; return $longest; }

UPDATE: There may be a minor bug in my use of POSIX, however I may be able to get round this as substr() automagically rounds down fractional values which are passed to it.

UPDATE^2: Fencepost error in ceil() function call fixed.

UPDATE^3: The reason why it fails is that the ordering can sometimes be wrong if there is more than one palindrome in the same string, due to the randomisation as the hash keys are fetched. I'll fix this at some point in the near future...

UPDATE^4: Issue fixed; Was in fact due to boneheadedness on my part. Now should work nicely, although not as pretty as it was.

UPDATE^5: Prettified a bit.

Elgon

"Stercus! Dixit Pooh. Eeyore, missilis lux navigii heffalumporum iaculas. Piglet, mecum ad cellae migratae secundae concurras."

Replies are listed 'Best First'.
Re^2: Finding longest palindrome from a string
by cLive ;-) (Prior) on Aug 13, 2004 at 19:48 UTC
    And so you should be :) Here are the benchmarks...

      Unfortunately, the benchmark doesn't tell the whole story. Many of the regex-based solutions fail on unescaped regex characters.

      All of the faster ones fail to correctly find the longest string when fed either of:

      1111111121111111111112111111111111111111111112111111111111211111 ababacababacadacababacadaeadacabaz

      (Including mine, which is sad, as the first one is my publish test string:()


      Examine what is said, not who speaks.
      "Efficiency is intelligent laziness." -David Dunham
      "Think for yourself!" - Abigail
      "Memory, processor, disk in that order on the hardware side. Algorithm, algorithm, algorithm on the code side." - tachyon