This program will search a file specified on the command line for any strings that contain all of "aeiou" in arbitrary places. It parsed a 12.3MB ascii text file matching 17000+ lines in 5 seconds.
#!/usr/bin/perl
if (<>) {
    @dog = <>;
    foreach $cat (@dog) {
        if ($cat =~ /a/i && $cat =~ /e/i && $cat =~ 
/i/i && $cat =~ /o/i && $cat =~/u/i) {
            print "$cat";
                                 }
                }
}

[download]

Replies are listed 'Best First'.
Re: vowel_search
by CheeseLord (Deacon) on Aug 17, 2001 at 04:09 UTC

    Your code at present will disregard a matching string on the first line. In addition, you read in the entire file at once, sucking up a good deal of memory if the file is quite large. Here's another way of doing it:

    #!/usr/bin/perl -w
use strict;  # Or don't. See if I care.

while (<>) {
   print if (/[Aa]/ && /[Ee]/ && /[Ii]/ && /[Oo]/ && /[Uu]/);
}

    [download]

    I only used the character classes there because (I hear) they're faster than /i matches. Could be wrong on that count, though.

    His Royal Cheeziness

[reply]
[d/l]
Re: vowel_search
by $code or die (Deacon) on Aug 17, 2001 at 06:13 UTC
    Not sure I'd want to load 17000 lines into an array in one go. I'd rather do something like this:
    #!/usr/bin/perl
while (<>) {
    if (/a/i && /e/i && /i/i && /o/i && /u/i) {
        print;
    }
}

    [download]
    update: Ahh, CheeseLord gave similar advice. Should have read all replies before posting :)

    Error: Keyboard not attached. Press F1 to continue.
[reply]
[d/l]
      Thank you $code or die, Your optimization took my total parse time from 5.2 seconds with my code to 2.9 seconds. I bow to my perlmonk teachers.
[reply]
Re: vowel_search
by runrig (Abbot) on Aug 17, 2001 at 04:08 UTC
    I wonder how this would benchmark since it doesn't rescan any part of the string:
    use strict;

my $vowels="aeiou";

for my $str (qw(abcdefghijklmnopqrstuvwxyz missing_vowels)) {
  my $tmp_v = $vowels;
  LOOP: {
    if ($str =~ /\G[^$tmp_v]*([$tmp_v])/i) {
      $tmp_v =~ s/$1//i;
      print "Succeeded: $str\n" and last LOOP unless $tmp_v;
      redo LOOP;
    }
    print "Failed: $str\n" and last LOOP;
  }
}

    [download]
    Update:Well, its dog slow. If there's a ton of consonants at the beginning of the string, it might be worthwhile, but otherwise, forget it. Oh well, it was worth a try :)
[reply]
[d/l]

Back to Cool Uses for Perl