G'day amma,

Welcome to the monastery.

I had a few issues with your example.

• $var2 begins with a '(' (left parenthesis) but that leading character is not present in the original string at that position.
• Should tlp_add_p0(0)) have an additional trailing ')' (right parenthesis) to match the others?
• Using etc is a bit vague so I'm guessing here. A better way to show this sort of thing is with an ellipsis, e.g. "A, B, C, ..., Z".
• Is the pattern 'NOT(tlp_p0(N) XOR tlp_add_p0(N))' common throughout or are there other formats?

If the repeated, internal pattern is the same throughout, you can do something like this:

$ perl -Mstrict -Mwarnings -E '
    my $x = qq{tlp_p2 <= (((((((NOT(tlp_p0(0) XOR tlp_add_p0(0)) AND\n
+};
    $x .= q{            NOT(tlp_p0(1) XOR tlp_add_p0(1))) AND  };
    $x .= q{NOT(tlp_p0(2) XOR tlp_add_p0(2))) etc};
    say for $x =~ /(NOT\(tlp_p0\(\d+\) XOR tlp_add_p0\(\d+\))/gm;
'
NOT(tlp_p0(0) XOR tlp_add_p0(0)
NOT(tlp_p0(1) XOR tlp_add_p0(1)
NOT(tlp_p0(2) XOR tlp_add_p0(2)
[download]

If you want to use split, I'd chop off any leading and trailing dross then split something like this:

$ perl -Mstrict -Mwarnings -E '
    my $x = qq{tlp_p2 <= (((((((NOT(tlp_p0(0) XOR tlp_add_p0(0)) AND\n
+};
    $x .= q{            NOT(tlp_p0(1) XOR tlp_add_p0(1))) AND  };
    $x .= q{NOT(tlp_p0(2) XOR tlp_add_p0(2))) etc};
    $x =~ s/\A\w+ <= \(+(.*?)\) etc\z/$1/ms;
    say for split /\)\s*AND\s+/ => $x;
'
NOT(tlp_p0(0) XOR tlp_add_p0(0)
NOT(tlp_p0(1) XOR tlp_add_p0(1))
NOT(tlp_p0(2) XOR tlp_add_p0(2))
[download]

This is a classical CS problem of infix parsing. See

• File:Abstract_syntax_tree_for_Euclidean_algorithm.svg
• Operator-precedence_parser
• Shunting-yard algorithm

This type of a problem is often presented as parsing infix notation for a 4 function calculator with parenthesis. That might give you a starting point.

I took this problem as inspiration to write a little script to visualize matching parantheses. I probably got carried away and it is not helping on the original problem.

use strict;
use warnings;
my $tlp_p2 = "(((((((NOT(tlp_p0(0) XOR tlp_add_p0(0)) AND NOT(tlp_p0(1
+) XOR tlp_add_p0(1))) AND  NOT(tlp_p0(2) XOR tlp_add_p0(2)))";

my $level  = 0;
my $tab    = "| ";
my %action = (
        '('       => sub { print "\n", $tab x ++$level, shift },
        ')'       => sub { print "\n", $tab x $level--, shift },
        'default' => sub { print shift },
        );

( $action{$_} // $action{'default'} )->($_) for $tlp_p2 =~ /./g;
[download]

The last line reads each character of the string $tlp_p2 individually and calls the corresponding sub from hash %action.

• ** Eternal question of parsing parentheses
• Re^2: parsing XML fragments (xml log files) with... a regex
• recursive algorithm for nested data structures
• balanced parens regexp hangs - solved
• A regexp to parse nested brackets containing strings
• Elegant examples to parse parenthesised strings
• Regexp::Common balanced curlies substitution
• RFC: A walkthrough from JSON ABNF to Regexp::Grammars
• How can I hope to use regular expressions without creating illegible and unmaintainable code?
• perlfaq6#Can I use Perl regular expressions to match balanced text?
• Re: Regex priorities

If you do not care about balanced parantheses it is simple:

use strict;
use warnings;
my $tlp_p2 = "(((((((NOT(tlp_p0(0) XOR tlp_add_p0(0)) AND
                                         NOT(tlp_p0(1) XOR tlp_add_p0(
+1))) AND  NOT(tlp_p0(2) XOR tlp_add_p0(2)))";
my @var = split /\s*AND\s*/, $tlp_p2;
$,="\n";
print @var;
[download]

BUT: already your proposed split for $var1 and $var2 does not have balanced parantheses. In fact, the closing ) at the end of $var2 corresponds to the opening ( at the beginning of $var1 if I have counted correctly. How do you want to deal with that situation?

Remark: If the terms in your equation are always of this type, you can create them bottom-up easily instead of extracting them from the equation...