10g Seeing Non-ASCII Characters as Multibyte

Hi all, this is probably a stupid question for most of you but I'm a developer and not a DBA by any means. We moved from 9i to 10g and we're having new issues. 9i, JavaScript and ColdFusion all see this string as a 30-character string:

“hello, this is just a ½ test”

Note that it uses non-ASCII slanted quotes and a non-ASCII symbol for "1/2" -- this is how MS Word would write this string.

Say, for example, that we have a VARCHAR2(30) column and I need to allow the application users to add up to 30 characters. They write it in Word, and Word says it's 30 characters. I test it using JavaScript and it's 30 characters. I use ColdFusion to make sure that I'm only handing off a 30-character string to Oracle and ColdFusion says it's 30 characters.

Oracle 10g sees it as 35 characters and the operation fails.

Our character set is AMERICAN_AMERICA.AL32UTF8 in 10g. We have lots of data that was stored in 9i with the AMERICAN_AMERICA.US7ASCII character set. Is it just a character set difference? What's our issue, here?

I'm basically looking for something I can tell our DBA team so we can get this fixed for our needs. What should I tell them? What would you do?

Thanks for the help!

Hi,
the issue is obvious - the column has capacity 30 bytes and the unicode string with non-ASCII characters is longer. Tell your DBA team to change the capacity of the column to 30 characters. That might be something like that:

Code:

---

alter table the_table modify the_column varchar2(30 char);

---

Thanks, I'll pass that info along to the team. It's definitely not obvious to me. Armed with that information I just found http://www.oracle-base.com/articles/...lization9i.php which really helps explain it all. Hopefully that does it!

Quote:

---

Originally Posted by Glowball

Hopefully that does it!

---

Absolutely!