Good point -- thanks. Though I might prefer the "publicly supported" method for doing that "conversion" (turning on the utf8 flag) -- one that doesn't involve a function whose initial underscore in the name asserts that it is intended to be "private" within Encode.pm (hence subject to change in a later release):

use Encode;
...
my $utf8_string = decode( 'utf8', $field_value_from_db );
[download]

Another point I should have considered earlier is that the OP might not want utf8 to be stored in the database (because of its "extra bulk"). If so, it would be better for the perl code that loads data into mysql to apply the appropriate encoding first (assuming the OP knows which non-unicode character set is suitable) -- e.g.:

my $field_value_for_db = encode( 'iso-8859-1', $utf8_string );
[download]

This way, nothing special needs to be done when pulling the data back out of the database for a non-utf8 display.