Re: Characterset problem with Oracle

• Post for FREE via your newsreader at post.usenet.com ****

"Peter Sylvester" <peters_no_spam_please_at_mitre.org> wrote in message news:ci4noi$5d3$1_at_newslocal.mitre.org... [...]
> Would you care to expand upon the overhead issue associated with UTF8.

Hi Peter

Unicode character sets needs more space to store data. Therefore all operations are a little slower.
Even Oracle clearly document this problem.... at http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96529/ch5.htm#1013253 you can read:
"If performance is your main concern, then consider using a single-byte database character set and storing Unicode data in the SQL NCHAR datatypes. Databases that use a multibyte database character set such as UTF8 have a performance overhead."

As you will never store short strings in CLOB instead of VARCHAR2, you should not use a Unicode character set if you don't need it.

> I would expect that with Java and other apps that deal with UTF natively
> that there might be less overhead, overall (client + server).

There are many Unicode encodings... Java, AFAIK, use UCS-2, i.e. a completely different encoding compared to UTF8!

Chris

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

• Usenet.com - The #1 Usenet Newsgroup Service on The Planet! *** http://www.usenet.com Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=