Unicode general questions
argo...@my-deja.com
We are currently using Oracle 7.3.4, but will
be migrating to 8i.
We are going to create a new database that must
support a number of languages, eventurally including
Chinese, Japanese and Korean. I'm thinking that Unicode
will be the way to go.
I notice that some Chinese charactersets used
in Oracle are 32 bits (4 bytes), and others are
only 16 bits (2 bytes).
However, Unicode is only 24 bits. To me, this
would limit the Chinese to the simplied 16 bit.
Is this right?
The other question is, if I use Unicode,
I can store in Japanese, Chinese, Korean,
English, German, etc., correct? It's the
application that will need to be able to handle
the display of the languages to screen, correct?
Are there any considerationsn on the Unix side?
Can I put Unicode on top of an English Unix Sun server,
and still have everything work?
Thanks,
Argosy
Sent via Deja.com http://www.deja.com/
Before you buy.
Jim Gregory
defined in bytes, not characters. If the Unicode encoding value ( variable
from 1 to 3 bytes per character/ideograph ) needs 3 bytes per, a
varchar2(12) is only big enough to store 4 characters not 12 and a 2 byte
character could only stuff 6 characters into a 12 byte field. Just make
sure that your field defs have enough bytes in them to store the maximum
number of characters needed.
--
Jim Gregory
Principal Consultant for Keane, Inc.
Currently assigned to NCR
"Opinions are my own and do not reflect
those of Keane or my clients"
<argo...@my-deja.com> wrote in message news:8bdl33$83l$1...@nnrp1.deja.com...
Galen Boyer
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.
Galen Boyer
Senior Database Architect
Primix Solutions
Matt
We're considering moving to a unicode database character set,
and have been testing on 8i for win98. To insert data in the
relevant language we changed the keyboard settings to the
required language. This worked fine using varchar's. Do the
nvarchar's come into play when a third language is required??
* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!
Christophe Galerne
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 cgal...@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