Re: Longest String Matching

[kcott]

G'day Dr Manhattan,

I think you decided that splitting the word into individual characters was a good approach and then got rather bogged down attempting to wring a solution from this technique. A regex using alternation would probably be a better way to do this.

$ perl -Mstrict -Mwarnings -E '
    my %lexicon = map { $_ => 1 } qw{own known};
    my $string = q{unknown};
    my $alt = join q{|} => sort { length $b <=> length $a } keys %lexi
+con;
    my $re = qr{ ( (?> $alt ) ) $ }x;
    $string =~ $re;
    say $1;
'
known

Also, unless the hash %lexicon already exists for some other purpose, you'd probably be better off with an array:

$ perl -Mstrict -Mwarnings -E '
    my @lexicon = qw{own known};
    my $string = q{unknown};
    my $alt = join q{|} => sort { length $b <=> length $a } @lexicon;
    my $re = qr{ ( (?> $alt ) ) $ }x;
    $string =~ $re;
    say $1;
'
known

Note how sort puts the longest string first, i.e. to test for "known" before testing for "own". See perlreftut for an introduction to alternation; and perlre for more information as well as a description of the (?>pattern) construct (in the Extended Patterns section).