Still not sure what you are doing here. But I tried to make some simplifications in the parsing part. I didn't really look in any detail at the DB part. I think you just need a HoA (Hash of Array). Each word would have a list of line numbers that it appeared on. A "unique" word would have just one line number in its array. Past that I'm not sure what this app is trying to do.

#!/usr/bin/perl

use strict;
use warnings; 
use DBI;

my $searcher = "owner.manager, owner.owner, owner.manown";

my $dbh = DBI->connect('DBI:mysql:db;host=localhost',
                        'user', 'pass')
                or die "Database connection: $!";

open( FILE, '<', $ARGV[0] ) || die "unable to open $ARGV[0]!";

my $line_num =0;
my %words = ();   #hash of array, list of line numbers per 
                  #each word

while ( my $line = <FILE> ) 
{  
    $line_num++;
    
    foreach my $word (split (/\s+/, $line))
    {
       $word =~ tr/^[a-zA-Z-]//; 
       $word =~ s|(\'.*)\b||g;  #maybe ?? not sure: what's->what?
       push (@{$words{$word}}, $line_num);
    }
}

# for every word, do a search...
# a "unique word" would be one who's array of line numbers has
# exactly 1 number
# I don't think that is what you wanted?
# I think what was meant is do one search per each word?

foreach my $word (%words)
{
   # @{words{$word}} is array of line numbers that word appeared 
   #   upon
   # line number appears twice if word appeared twice on that 
   #   line
   ... do something in DB with $word....
}

# $uword { '$seen{ $word }[ 1 ]' }  - should be the line(s) of test th
+e search string came from

#no, that would be @{words{$word}}
[download]