Re: Changing Character Set

Kalle wrote:
> Hi all,
>
> I wonder if it is possible to just on the fly change the character set
> lie from WE8DEC to UTF8 on 9i.

You mean of the database? It's hardly an "on the fly" change for the database... more like a 'get to the MOUNT state and issue an alter database command...'.

And I don't think you should use the UTF8 character set, because it adheres to an older Unicode standard, its use is therefore officially deprecated, and it's supplied for backwards compatibility only. AL32UTF8 is an acceptable Unicode variable-width character set.

> I have run the csscan and it pointed some columns must be lenghtened,
> after I have done this, is it safe to alter the character set, or is
> there something else I need to consider...

Csscan only reports truncation and data loss/corruption errors. It doesn't, for example, report character set mismatch errors (though you're unlikely to suffer from them). So you could still face problems, but there's not much you can do about it one way or another... though you might want to practice on a cloned subset of data first, before letting rip on the production box.

Otherwise, if csscan reports no errors, you should be good to go. The only residual stuff to watch out for is, for example, altered sort orders as a result of the change... but I again doubt you'd encounter any of that. But that's why we test things first, to make *sure* you'd not encounter it!

So no, therefore: there's no 'Oh, let's just alter the database character set' mode. Like anything else that potentially modifies every bit of data in your database, you treat it with extreme caution.

Regards
HJR
>
> Thank you again for your valuable help!
>
> Kalle
>
Received on Tue Nov 30 2004 - 22:58:10 CST