Database Software

I am not clear: is the problem with unicode metadata ,i.e. the columns are named in some unicode language? Then you have to run the Repository Assistant again and select "Add More Languages to Support Object Translations" These are the languages in which you want to edit and display the logical names and descriptions of objects created in OWB

Or is your issue with unicode data in the source? Then is the source an Oracle database with characterset UTF-8? And does the target database have the same characterset?

Nikolai Rochnik

Nikolai,
My problem is with data only.
i have oracle database with characterset UTF-8 as a source and target.
i have few tables in source schema where i have used unicode datatypes (NVARCHAR2,NCHAR etc)but when i try to import those tables into the OWB mappings those columns (unicode)are missing. so the result is ....i can't able to fetch the data from these columns.
in short if still i am not clear .."is there any special handling in OWB if Unicode is found in any of the fields of the tables?"

Thanks

Ok, got it - the problem is that in the current release NVARCHAR2, NCHAR datatypes are not supported. You are actually notified of this explicitly when you perform the metadata import. From the summary at the end (where column NAME is of type NVARCHAR2):
"The following actions have been applied:

+ Table
+ TEST_TABLE ................... : Created with Second Class Object Failure
+ Columns
+ ID ........................... : Created
+ NAME ......................... : Skipped (unsupported datatype)"

So what do you do?

1. Wait till our new upcoming release (Nov 2004) that adds support for ALL Oracle datatypes, including National Character types. Just to be sure I tried importing a table with NVARCHAR2 type with the beta version of our new release and everything works as expected.

2. Use a workaround for now to write views to translate currently unsupported NVARCHAR2, NCHAR types into supported types: VARCHAR2, CHAR, etc.

Nikolai Rochnik