question marks on insert of Unicode Japanese text with Perl

[Dan McCormick]

Hi,

When I try to insert Unicode Japanese text into a column, it gets converted to question marks. I have a test table, t, with one column, kw:

mysql> select hex(kw), hex("?") from t;
+---------+----------+
| hex(kw) | hex("?") |
+---------+----------+
| 3F3F3F | 3F |
+---------+----------+

The table was created to support UTF-8:

create table t ( kw varchar(255) character set utf8 collate utf8_unicode_ci NOT NULL default '' ) ENGINE=MyISAM DEFAULT CHARSET=utf8;

I'm inserting via a Perl script. I first set the character set:

$dbh->do("set character set utf8");

I then get a string from a web site, and Devel::Peek::Dump assures me that it's Unicode:

SV = PV(0x953b584) at 0x91309f8
REFCNT = 1
FLAGS = (PADBUSY,PADMY,POK,pPOK,UTF8)
PV = 0x94f8c80 "\343\202\253\343\203\251\343\203\274"\0 [UTF8 "\x{30ab}\x{30e9}\x{30fc}"]
CUR = 9
LEN = 12

I can also use Unicode::CharName to look at each of the three characters that are returned:

KATAKANA LETTER KA
KATAKANA LETTER RA
KATAKANA-HIRAGANA PROLONGED SOUND MARK

I then do a simple insert:
my $sth = $dbh->prepare("insert into t (kw) values (?)");
$sth->execute( $t );

and end up with the set of three question marks shown above.

I'm using mysql 4.1.13, Perl 5.8.5, and DBD::mysql 3.0002.

Any advice would be greatly appreciated.

Thanks,
Dan