#!/usr/bin/perl

use strict;
use warnings;

use Benchmark;

my @sizes = (10, 25, 50, 100, 250, 500, 1000);
my @chars = ('A' .. 'Z', 'a' .. 'z', 0 .. 9);

our @d = map {join "" => map {$chars [rand @chars]} 1 .. $_} @sizes;

map {
    Benchmark::cmpthese timethese (-2 => {
        "simple_$sizes[$_]"   => '$_ = $::d[' . $_ . ']; s/\D//g;',
        "multiple_$sizes[$_]" => '$_ = $::d[' . $_ . ']; s/\D+//g;'
    }, 'none');
} 0 .. $#sizes
__END__
                Rate   simple_10 multiple_10
simple_10   196495/s          --        -15%
multiple_10 231225/s         18%          --
                Rate   simple_25 multiple_25
simple_25    89788/s          --        -50%
multiple_25 180650/s        101%          --
                Rate   simple_50 multiple_50
simple_50    47507/s          --        -64%
multiple_50 130727/s        175%          --
                 Rate   simple_100 multiple_100
simple_100    23206/s           --         -77%
multiple_100 103096/s         344%           --
                Rate   simple_250 multiple_250
simple_250   10488/s           --         -71%
multiple_250 36407/s         247%           --
                Rate   simple_500 multiple_500
simple_500    5046/s           --         -75%
multiple_500 20382/s         304%           --
                 Rate   simple_1000 multiple_1000
simple_1000    2528/s            --          -76%
multiple_1000 10549/s          317%            --
[download]

Abigail

In this case transliteration is really the most efficient solution though. Consider the results of adding         "xlit_$sizes[$_]"     => '$_ = $::d[' . $_ . ']; tr/0-9//cd;', to the benchmark:

                Rate   simple_10 multiple_10     xlit_10
simple_10    86400/s          --        -31%        -70%
multiple_10 124615/s         44%          --        -57%
xlit_10     292712/s        239%        135%          --
                Rate   simple_25 multiple_25     xlit_25
simple_25    45324/s          --        -49%        -82%
multiple_25  88062/s         94%          --        -65%
xlit_25     248802/s        449%        183%          --
                Rate   simple_50 multiple_50     xlit_50
simple_50    23823/s          --        -71%        -89%
multiple_50  82566/s        247%          --        -62%
xlit_50     218684/s        818%        165%          --
                 Rate   simple_100 multiple_100     xlit_100
simple_100    13397/s           --         -69%         -92%
multiple_100  43191/s         222%           --         -74%
xlit_100     168434/s        1157%         290%           --
                 Rate   simple_250 multiple_250     xlit_250
simple_250     5608/s           --         -71%         -95%
multiple_250  19639/s         250%           --         -81%
xlit_250     103656/s        1748%         428%           --
                Rate   simple_500 multiple_500     xlit_500
simple_500    2832/s           --         -72%         -95%
multiple_500 10189/s         260%           --         -83%
xlit_500     59072/s        1986%         480%           --
                 Rate   simple_1000 multiple_1000     xlit_1000
simple_1000    1380/s            --          -77%          -96%
multiple_1000  5939/s          330%            --          -83%
xlit_1000     34457/s         2397%          480%            --
[download]

Esp in large data sets, transliteration screams.

Makeshifts last the longest.

True, but my point was the pattern of s/PAT//g, which would benefit to be written as s/PAT+//g. tr isn't as flexible - not even in this case. \D follow the locale and Unicode rules when appropriate, where as the tr has the digits hardcoded.

Abigail

Ok, thanks. Also did have one more question about this -- I didn't take into consideration about valid non-numeric (only decmial). So for example, 1a2b3.4c5d6e, I would want 123.456, not just 123456 -- tried a few combinations, but nothing yet. I'm missing something obvious I know ;) Any pointers?

It always helps if you are specific. Noone enjoys a game of "How do I do X?", "This is how you do X", "But I don't really want to do X, I want to do Y".

So, from 1a2b3.4c5d6e, you want 123.456. But what if you have 1a2b.3c4d.5e6f? What do you want then?

Abigail

If I understand you correctly, you want to remove floating point numbers as well, together with normal integer numbers but not if they are part of a normal word (like 'HAL1'), don't you?

Then maybe you should have a look at Regexp::Common. Together with putting whitespace around the regex you should get it to work.

-- Hofmator

Basically, in any given string, I want only numeric and decimal values retained. So, basically:

1a => 1
1a.2b => 1.2
1a.2b.3c => 1.2.3

^^ Just as those show, no matter how many decimals or numeric values, those are the only characters I want retained. Hope this clears it up just a bit.

So, you want to delete anything that isn't a number or a dot? Just do:

    s/[^\d.]+//g;
[download]

or

    tr/0-9.//cd;
[download]

Abigail