Yes, well, that also failed ... this is not a good idea ... :-)
First, to support UTF-8 in Oracle, you want to use the character set "AL32UTF8".
Is this a new database? Or is the existing database with data already in it?
If we are talking about a new database, you just need to drop the database and create it again by specifying the AL32UTF8 character set in DBCA when creating the database.
If it is an existing database with existing data, it is a little more complicated. What character set are currently used? Is AL32UTF8 a direct extension of this character set?
This is a truly nontrivial subject. In short, you need to know the current character set and the character set you are accessing, and you need to run the "CSSCAN" utility to check for compatibility and data conversion problems.
The whole story you should check the following document on MOS:
Change NLS_CHARACTERSET to AL32UTF8 / UTF8 (Unicode) [ID 260192,1]
Hope this helps.
source share