Hi. A little detective work lead me to your previous posting which supplies some context for your question here. Yes, DBD::CSV will be slow with a file as large as the one you are dealing with. The fastest way to get your data from a CSV file into a quickly-searchable form is to use the loading mechanism of a database (for example LOAD INFILE with MySQL). Using a database would also simplify and speed up future updates and searches. If you absolutely must use BerekeleyDB instead of a database, then you can convert from CSV to BerekeleyDB with something like this:

#!/usr/bin/perl
use warnings;
use strict;
use BerkeleyDB;

my( $csv_file, $berk_file) = qw( dict.txt dict );
my $db = BerkeleyDB::Hash->new(
    -Filename => $berk_file,
    -Flags    => DB_CREATE
) or die "Cannot open file '$berk_file': $! $BerkeleyDB::Error\n";
open( DICTE, $csv_file )
  or die "Cannot open file '$csv_file': $!\n";

for (<DICTE>) {
    chomp;
    my($key, $value) = split(/;/,$_,2);
    $db->db_put($key,$value);
}

# the file "dict" is now a BerkeleyDB file with entire
# contents of the CSV file "dict.txt"
[download]