I set nls_length_semantics = 'CHAR' for Oracle 9i database, the character set is 'AL32UFT8', I assumed it stored characters based on number of Unicode unit instead of number of bytes base on the Oracle menu, however it seems not working when I try to load Japanese into it.

Create table test (
Lookup_id number(4),
Name nvarchar2(240));

It allowed me to insert 80 Japanese characters into test.name, but it will fail over 80 Japanese characters.
Please help.