in reply to Matching data against non-consecutive range

I do not think that I would ever use a regex to match specific numbers in a given range. I would use a lookup table instead.

#!/usr/bin/perl
use strict;

# Zone 8
my %z8;
$z8{$_}++ for ( 10..374, 376..379, 382..385, 388..499, 530..534, 541..
+543, 618, 619, 700..704, 707..709 );

my $num = "34";
print "Found it!\n" if $z8{$num};

[download]

Another option would be to pull the zone information using Business::UPS. Store it in a database or use Storable's freeze and thaw to save and load your data. You'll probably also want to determine how often to automatically freshen your information.

Replies are listed 'Best First'.
Re^2: Matching data against non-consecutive range
by bart (Canon) on Jan 27, 2005 at 22:07 UTC
    Using Regex::PreSuf, you can convert that list of numbers into a regex (or at least, a string you can use as a regex). Like this: 
    use Regex::PreSuf;
my $re = presuf(10..374, 376..379, 382..385, 388..499, 530..534, 541..
+543, 618, 619, 700..704, 707..709);
print "Regex: /$re/\n";

    [download]
    This prints: 
    Regex: /(?:1(?:0[0123456789]|1[0123456789]|2[0123456789]|3[0123456789]
+|4[0123456789]|5[0123456789]|6[0123456789]|7[0123456789]|8[0123456789
+]|9[0123456789]|[0123456789])|2(?:0[0123456789]|1[0123456789]|2[01234
+56789]|3[0123456789]|4[0123456789]|5[0123456789]|6[0123456789]|7[0123
+456789]|8[0123456789]|9[0123456789]|[0123456789])|3(?:0[0123456789]|1
+[0123456789]|2[0123456789]|3[0123456789]|4[0123456789]|5[0123456789]|
+6[0123456789]|7[012346789]|8[234589]|9[0123456789]|[0123456789])|4(?:
+0[0123456789]|1[0123456789]|2[0123456789]|3[0123456789]|4[0123456789]
+|5[0123456789]|6[0123456789]|7[0123456789]|8[0123456789]|9[0123456789
+]|[0123456789])|5(?:3[01234]|4[123]|[0123456789])|6(?:1[89]|[01234567
+89])|7(?:0[01234789]|[0123456789])|8[0123456789]|9[0123456789])/

    [download]

    Wow, that's big: length($re) is 746 bytes. It won't hurt though, you can just use it to match: 

    $zip = '34';
if($zip =~ /\b$re\b/o) {
    print "Got a match for $zip\n";
}

    [download]
    which prints: 
    Got a match for 34

    [download]

    Considered by jmanning2k - Break Long Line
    Unconsidered by castaway - Keep/Edit/Delete: 7/7/0 - Get a working browser, see Re^2: Monastery Gates page is too wide (causes)

    Note by bart: I use Firefox as my standard browser, and it renders correctly for me. I asked on the Chatterbox, and both castaway and corion said it looked normal to them, too. They suggested it could be caused by a difference in site settings for wrapping... *shrug* I don't know any more.

[reply]
[d/l]
[select]

      IMO its a good idea to attach "don't do this" notices to a solution like this. Its hopelessly inefficient in comparison to the hash lookup.

      ---
      demerphq

[reply]
        Have you benchmarked it? I haven't... it might not be that bad, and it's most likely a lot better than just 
        $re = join '|', @list;

        [download]
        anyway...

        I could try and optimize it a little by making sure it doesn't even attempt to match if it doesn't see a digit: 

        /\b(?=\d)$re\b/o

        [download]

        Ideally, I'd think this would be an excellent problem to solve using a regexp assertion in perl code, thus matching all digits first and then testing if it's within range with a hash, all inside the regexp — if only assertions weren't that hard to implement.

        p.s. grinder has yet another alternative solution to using Regex::PreSuf, it's currently up on his scratchpad, but for some otherwise noble but in this case silly reason of policy (because he frontpaged the thread) he doesn't dare to post it. Please urge him to post it, it'd be a waste to the site if he didn't. He won't listen to me. :)

[reply]
[d/l]
[select]
        It's probably time-efficient enough for most purposes. And the hash lookup is probably memory-efficient enough for most purposes.
[reply]

In Section Seekers of Perl Wisdom