Split String Based on RegEx in Perl

[thehemi]

Let's assume this is what I'm working with...

$string = 'xxxxxxxxyyyy';
$regex = 'xxxxxxxx';

And my goal is to have the following results...

$value1 = 'xxxxxxxx';
$value2 = 'yyyy'; (Remainder after $regex)

What commands am I using to get to that point?
I have used regex to remove and/or alter text, but
I have never used them to split apart a string.

I can hack around to get a result, but I prefer to
do it as 'properly' as possible and minimizing lines
and execution cycles.

[Cabhan]

Alrighty. So Perl has a built-in split function that is what you are going to use:
http://perldoc.perl.org/functions/split.html

Now, the only trick is that split takes a delimiter, whereas you want to actually tell it the first bit, and have the rest split. You can do this by using a special Perl regex feature: the positive lookbehind. For your particular example, for instance, the code would be:

($value1, $value2) = split /(?<=xxxxxxxx)/, $string;

What this does is split $string whenever it reaches a section immediately following the string 'xxxxxxxx'. That is, the regex looks backward and says "Do I see the string 'xxxxxxxx' just behind the current position? If so, then I match the CURRENT position."

This is one of the more advanced features, and is described in the Camel book under the "Fancy Regular Expressions" section.

[thehemi]

Very great advice, thanks for the information!
However, I don't think it likes using a regex...

-bash-3.00$ ./script.sh
Variable length lookbehind not implemented in regex

[Cabhan]

Aha. Are you using exactly what you posted here, or are you doing something different? Unfortunately, lookbehinds cannot be variable-length: they cannot use quantifiers. For instance, this works properly:

alex@danu ~/test/perl $ cat using_x_in_regex 
#!/usr/bin/perl -w

use strict;

my $string = 'xxxxyyyy';

my($val1, $val2) = split /(?<=xxxx)/, $string;

print $val1, "\n", $val2, "\n";
alex@danu ~/test/perl $ ./using_x_in_regex 
xxxx
yyyy

This does not:

alex@danu ~/test/perl $ cat using_x_in_regex_wrong 
#!/usr/bin/perl -w

use strict;

my $string = 'xxxxyyyy';

my($val1, $val2) = split /(?<=x{2,4})/, $string;

print $val1, "\n", $val2, "\n";
alex@danu ~/test/perl $ ./using_x_in_regex_wrong 
Variable length lookbehind not implemented in regex; marked by <-- HERE in m/(?<=x{2,4}) <-- HERE / at ./using_x_in_regex_wrong line 7.

[thehemi]

Are you using exactly what you posted here, or are you doing something different?

Sorry if my post wasn't clear. I'm using a regular expression.
Something along the lines of http://[a-z]+\.domain\.[a-z]{3}.

[Cabhan]

Gotcha. This isn't a call for splitting at all. Just use matching.

my($sub, $domain, $end) = $string =~ m|http://([a-z]+)\.(domain)\.([a-z]{3})|;

If you want to just split the string at some point, then just do two matches:

my($val1, $val2) = $string =~ m|(http://[a-z]+\.)(domain\.[a-z]{3})|;

See?

EDIT:

As a side note, you can use lookbacks with regexes, just it needs to have a specific length. So for instance, if you replaced "[a-z]+" with a specific length. Of course, you can't do that in this particular problem, but lookbacks can still be very useful.

[thehemi]

That works PERFECT. Thanks for the help!