Hi all

I'm using Tie::RDBM with postgresql to store some data. My problem is that the values obrained back from the database does not have set unicode (utf8) flag and are treated as a octets instead of strings.

Yes, of course I can use $data = decode('utf8', $hash{'key'}) for each data (both keys and values) obtained from database, but... Is there better way to do this? Is there something like DBM filers (filter_fetch_key, filter_fetch_value)?

My testing code is here

#!/usr/bin/perl_parallel -w
# For Emacs: -*- mode:cperl; mode:folding; -*-
use strict;
use utf8;
use Tie::RDBM;
use Encode;

my %data;
my $db_name = 'unicode';
my $db_host = 'localhost';
my $db_user = 'elric';
my $db_pass = 'test01';

tie(%data, 'Tie::RDBM', {
 db         => "dbi:Pg:dbname=$db_name;host=$db_host;",
 user       => $db_user,
 password   => $db_pass,
 table      => 'Demo',
 create     => 1,
 drop       => 1,
 autocommit => 1,
 DEBUG      => 0
}) or die $!;

my $counter = 0;
open DATA, '<:utf8', 'input.utf8' or die $!;
while (<DATA>) {
  chomp;
  my ($key, $value) = split(':', $_, 2);
  $data{$key} = $value;
  # BAD - print octets
  print $data{$key}, "\n";
  # OK - print string
  print decode('utf8', $data{$key}), "\n";
  $counter++;
}
close DATA;

print 'number of keys witten: ', $counter, "\n";
print 'number of keys in database: ', scalar keys %data, "\n";

# BAD - print octets
print join(', ', keys %data), "\n";

# OK - print strings
print join(', ', map { decode('utf8', $_)} keys %data), "\n";

untie %data;
[download]

Thanks for your help