So, you have converted the original data from euc-jp to utf8, you've stored that in the database, you've fetched the utf8 data back from the database, and now you want to convert it back to euc-jp, but using the simple "encode" call for that last step was a bust. Did I get that right?

If so, the problem is probably that when you get the data back from the database, perl doesn't know that it's utf8 anymore, and the encode function will do the wrong thing as a result.

The stuff coming back from the database is a string of "octets" (which happen to constitute valid utf8 data in Japanese), and you need to convert from utf8 octets to euc-jp octets, using the "from_to" function from Encode (I'll tweak the variable names for clarity):

# prepare original data for the database:

$utf8_string = decode( "euc-jp", $euc_string );

# store that to the database, then some time later,
# fetch it back, and convert back to euc-jp:

from_to( $db_octets, "utf8", "euc-jp" );

# now, $db_octets should be readable in an euc-jp display.
[download]

Another way to do it, depending on whether you want to do other character-based things with the stuff that comes back from the database, is to go ahead and tell perl that the database string is utf8 (by "converting" it from utf8 octets to utf8 characters, if that make sense), and then just convert it to euc-jp when you print it -- e.g. to STDOUT:

$utf8_string = decode( "utf8", $db_octets );

binmode STDOUT, ":encoding(euc-jp)";

print $utf8_string;
[download]