Here's my shot at it:
#!/usr/bin/perl
use strict;
use warnings;
# set up
srand;
my @one = (1,2,3,4,5,6);
my @two = (2,4,6,8,9,12);
# the value that needs replacing
my $bad = 4;
# figure out where that value is in list one
### my $index = index(join('',@one),$bad); # <- bad code
my $index = 0;
my $found = 0;
foreach my $item (@one) {
if ($item == $bad) {
$found++;
last;
}
$index++;
}
die "The value '$bad' was not found in the list\n" unless ($found);
# find elements of list two that aren't already in list one
my $re = join('|',@one);
my @candidates = grep(! /^(?:$re)$/, @two);
# replace unwanted list one value with random selection from list two
$one[$index] = $candidates[rand @candidates];
use Data::Dumper::Simple;
print Dumper(@one);
__END__
@one = (
1,
2,
3,
8,
5,
6
);
Update: I realize that the line that generates $index is a potential point of failure via false positives (matching "10" when $bad = 1, for example). However, I've discussed this with the OP and he says it's not an issue, due to the fact that the values are all 4-digit numbers.
Update 2: The above statement no longer applies post-patch.
---
It's all fine and dandy until someone has to look at the code.