Assuming each RSno has only one StarName and vice versa, it would be simpler to use a hash for the conversion

#!perl
use strict;
my @Stars = qw(CYP2C19_10 CYP2C19_12); 
return if (@Stars == 0);

# name,no
my $ref_file = "test.csv";
open my $star_FH,'<',$ref_file
   or die "Could not find reference file $ref_file : $!";

my %RSno=();
my $count=0;
print "Reading from ref_file $ref_file .. ";
while (<$star_FH>){      
  s/\s+$//;
  my ($no,$name) = split ",",$_;
  $RSno{$name}   = $no;
  ++$count;
}
close $star_FH;
print "$count records read\n";

for my $StarIndex (@Stars){
  if (exists $RSno{$StarIndex}){
    print "$StarIndex is $RSno{$StarIndex}\n";
  } else {
    print "$StarIndex NO CONVERSION\n";
  }
}
[download]

That worked! Thanks! Can you explain what the difference was (and what the above code means?) =P Thank you so much!

s/\s+$//

demonstrates the use of the s command, which takes this form: s/replace stuff matching this pattern/with this/. In this case the "with this" is blank, which means it will just remove stuff matching the pattern. \s is an abbreviation for white space characters - space, tab, line feed. + means match one or more such characters in a row. $ means match only if the pattern is found at the very end of a string.

This is an inadequate explanation; see the Tutorials section, especially here.

Dum Spiro Spero

The csv file was saved under windows, yes. Do I need to open it a different way?

Ah. That explains it. Windows Linebreak consists of CarriageReturn and LineFeed while Linux uses a LineFeed. Your chomp only strips the LineFeed, but leaves the CarriageReturn in place.

You can either pre-process the CSV file with dos2unix / flip or some other tool or manually strip the CarriageReturn in perl.