The script below is based upon the Text::CSV documentation:

use Modern::Perl;
use Text::CSV;

my @Sample_Ids_m;
my $csv = Text::CSV->new( { binary => 1 } )    # should set binary att
+ribute.
  or die "Cannot use CSV: " . Text::CSV->error_diag();


while ( my $row = $csv->getline(*DATA) ) {
    push @Sample_Ids_m, $row->[6];            # Element 6 is the ID
}

say for @Sample_Ids_m;                         # Print the collected I
+Ds

__DATA__
John Doe,G building,05-Aug-2012,08-Aug-2012,"New York City, NY",ABC234
+5,UCD23467, 
John Moe,H building,05-Aug-2012,08-Aug-2012,New York City,DEF2345,UCD8
+0645,
John Slo,I building,05-Aug-2012,08-Aug-2012,New York City,GHI2345,UCD7
+6765,
John Hor,j building,05-Aug-2012,08-Aug-2012,New York City,JKL2345,UCD8
+7111,
[download]

Output:

UCD23467
UCD80645
UCD76765
UCD87111
[download]

Adapted for your csv file (untested):

use Modern::Perl;
use Text::CSV;

my @Sample_Ids_m;
my $csv = Text::CSV->new( { binary => 1 } )    # should set binary att
+ribute.
  or die "Cannot use CSV: " . Text::CSV->error_diag();

open my $fh, "<", "Merged_CSVfiles.csv" or die "Merged_CSVfiles.csv: $
+!";
while ( my $row = $csv->getline($fh) ) {

    push @Sample_Ids_m, $row->[16];            # Element 16 is the ID
}
$csv->eof or $csv->error_diag();
close $fh;

say for @Sample_Ids_m;                         # Print the collected I
+Ds
[download]

From your code above, it looks like the ID is the 16th element in @columns, and the $row->[16] notation grabs it after parsing a csv line. See if this may assist your scripting...