Martijn van Oosterhout wrote:
> I don't think making a special typlen value just for a type that can
> store a single UTF-8 character is smart. I just can't see enough use to
> make it worth it.
>
Assuming that we can set encoding per-column one day, I agree. If you
have a CHAR(1) field, you're going to store codes like 'A', 'B', 'C' or
'1', '2', '3' in it, and you don't need UTF-8 for that.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com