use Encode;
use DBI;  # (or whatever you use for your PostgreSQL)

my $dbstr; # suppose this holds "unicode" data from the DB

# ... do whatever it takes to fetch a value into $dbstr;
# since DBI might not be "unicode-aware", you may need to
# coerce perl into treating the value as unicode:

my $unistr = decode( 'utf8', $dbstr );
my $latin1str = encode( 'iso-8859-1', $unistr );

print $latin1str;

##</code><code>##

# snippet to convert utf8 to latin1 -- NB: only works for utf8
# characters that correlate to unicode \x{0000} - \x{00ff}
# (and you really should figure out how to convert using a module)

my @bytes = unpack C*, $_; # break utf8 string into bytes
$_ = '';

while ( @bytes ) {
    my $b = shift @bytes;
    if ( $b & 0x80 ) { # start of utf8 (latin1) character
        my $c = ( $b & 3 ) << 6;  # 1st utf8 byte carries top 2 latin1 bits
        $_ .= chr( $c | ( shift @bytes & 0x3f ));  # 2nd byte has the other 6 bits
    } else {
        $_ .= chr( $b );  # utf8 ascii is just ascii.
    }
}

# now $_ holds latin1 (single-byte, iso-8859-1) characters