Re: Unicode general questions

Galen Boyer wrote:
>
> I would hold off on this until you go to 8i. Double byte Unicode is not
> supported within Oracle with versions below 8, while an earlier version of
> UTF8 (I believe it is called UTF7) is supported in Oracle 7.3.
>
> In 8 or 8i, you will have the NCHAR and NVARCHAR fields which are there for
> unicode support. Plus, if you are using java, it works in UTF2 so you are
> covered completely. One thing to remember is that all your ascii data will
> now be doubled in size. Double byte for all, what used to single-byte,
> ascii. This could cause network problems, but since you will be sending
> across asian characters, you will probably already be ready to deal with
> the network issues.
>
> Now for UTF8, (still keeping in mind 8 only), you will need to bring your
> database up in UTF8 character set. Then, your char and varchar will be
> able to handle UTF8, but you may have to convert to UTF8 from whatever
> character set your application is in, even java would have to convert from
> UTF2 to UTF8.
>
> If you go with 7.3, bring your database up in AL24... (I don't remember
> exactly which one this is). Then, good luck getting your tools to send
> multibyte characters across the sqlnet layer successfully.

The right encoding is AL24UTFFSS. Oracle 8 and 8i support the character set UTF8. There is no problem sending multibyte characters across the sqlnet layer if you correctly set your NLS_LANG on the client side.

Christophe

Christophe L. Galerne                                           cgalerne_at_tumbleweed.com
Software Engineer                                                   Tel: (650) 216 2029
"I've been sitting here, just wasting time, drinking, smoking, trying to free my mind."
                                                                            -- Kid Rock