Converting database from US7ASCII to UTF8

I am on HP UNIX 11i using Oracle 9.2.0.5.

I had a database in database character set US7ASCII. I recently converted it to database character set UTF8. Since I do not have any nchar or nvarchar2 columns, national character set is irrelevant even though it was UTF8 to begin with.

Export was done using NLS_LANG=AMERICAN_AMERICA.US7ASCII and same setting was used for import.

During import some rows got rejected because column lengths of some varchar2 column will not fit UTF8 chacater set. I know UTF8 needs more bytes if character is outside of ASCII. My question is since I had US7ASCII database to begin with, why my rows got rejected, since UTF8 is a superset
of US7ASCII.

Looking at failed rows generated during import, I did notice that some rows had characters such as Ö (O with two .. on top), R with a trademark on top of it. If my database character set was ASCII, how this character was inserted in the first place. What is the meaning of US7ASCII if my database had these characters. All columns with error are varchar2 columns.

Appreciate your insight.

Thanks a lot. Received on Sun Dec 11 2005 - 13:06:16 CST