Removing duplicate fields with MARC::record

[Michael Bowden]

Hi Folks:
 
It is me again.  I have another question...
 
I am helping someone clean up her database.  Somehow, the 856 field in her MARC 
records has duplicated itself several times.  She has some records with 15+ 
duplicate 856 fields.  So I am trying, unsuccessfully, to modify one of my 
scripts to delete the duplicate fields.  The problem I am having is that out of 
the 15+ 856 fields, 3 of the 856 fields are unique to the record and need to 
stay.  Here is my code so far:
 
## create a MARC::File::USMARC object
 
use MARC::Batch;
 
my $infile = shift;
my $otfile = shift;
my $batch = MARC::Batc
 
Michael L. Bowden
Coordinator of Automation and Access Services
Associate Professor, Information Science
Harrisburg Area Community College
One HACC Drive
Harrisburg, PA 17110-2999

E: [EMAIL PROTECTED]
T: 717.780.1936
F: 717.780.2462h->new('USMARC',$infile);
$batch->strict_off();
 
open (DBOUT, "> $otfile");
## if $file isn't defined we had trouble with the file
## so exit
 
if (not($infile)) {
print $MARC::File::ERROR,"\n";
exit(0);
}
 
while (my $record = $batch->next()) {
   my @m856 = $record->field('856');
   @m856 = sort {$a cmp $b} @m856;
   my %seen = ();
   my @new856 = ();
   
   foreach ( $record->fields() ) {
      if (@m856) {
         foreach $f (@m856) {
             next if ($seen{ $f }++);
           push @new856, $f;
             $record->delete_field($f);
             }
         }
   }
   $record->insert_fields_ordered( @new856 );
   print DBOUT $record->as_usmarc();
}
 
When I run this script, It put ALL the 856 fields back in the record and they 
are not sort. What am I doing wrong?
 
TIA!
 
Michael

Harrisburg Area Community College