I used a similar idea but did as follows:

my $file = '...the OPs File example..." #used as a string
                                        # to simplify my
                                        # test
my @gi_ids = $file =~/gi\|(\d+)\|/g;
[download]

That leaves the list @gi_ids containing all of the gi id's that were in the string (i.e., it finds 4 gi id's in the OPs example).

Just another possible strategy. It looks to me like all of the respondents' ideas would work fine. Mine is just another variation.

Hope that helps.

ack Albuquerque, NM

You can use this code..,

use strict;
use warnings;
use Data::Dumper;
my($File,$FILE_FH,$data,@total_data,$new,@gi);
$File="./fasta.txt";
open  $FILE_FH, '<', $File or die "Failed to Open File $File for $!";
 $data=<$FILE_FH>;
 @total_data=split(/gi\|(\d+)\|/,$data);
foreach $new(@total_data)
{
    push @gi,$new if($new =~ /^\d+$/);
}
print Dumper @gi;
[download]