Greek Character Set Problem

Question

Dear All,
We are currently working on Oracle 8i to 9i migration and unicode implementation of an application SampleDB. We encountered an issue (described below) while the data migration.
Issue:
We have a database in oracle8i with
NLS_LANGUAGE AMERICAN
NLS_CHARACTERSET WE8ISO8859P1
NLS_NCHAR_CHARACTERSET WE8ISO8859P1
The character set WE8ISO8859P1 does not support GREEK language. The user inserts/modifies records in this database from the front-end. The front end installed in the client machine has the following setting:
Control Panel->Regional settings ->Advanced ->Language for non-unicode programs = Greek
Oracle client installed in English
Greek characters have been inserted/ updated into this database and retrieved and displayed correctly by the GUI.
Note: Greek data has been inserted into WE8ISO8859P1 database which does not support Greek characters.
Hence the actual data that is stored in the database are not Greek. Only while displaying, they are getting converted into Greek characters due to the settings in the client machine.
Requirement: Migrate the database to Oracle 9i - Unicode.
While migration, we need to eliminate any data discrepancy and make sure the data in the Unicode database is in Greek.
Can any one suggest us a method to do the migration to avoid the above mentioned problem?
Further, can any one tell us a way to verify what characters get stored in the database and help us identify if they are real Greek characters

Answer 1

The character set WE8ISO8859P1 does not support GREEK language.If you want to support Greek characters (or 'math' symbols), look at the EL8... char sets.
Control Panel->Regional settings ->Advanced
->Language for non-unicode programs = Greek
Oracle client installed in EnglishThis is dependent on the app and Windows. What is the NLS_LANG value in this Oracle home? NLS_LANG should be adjusted to fit the application.
Note: Greek data has been inserted into WE8ISO8859P1
database which does not support Greek characters.
Hence the actual data that is stored in the database
are not Greek. Only while displaying, they are
getting converted into Greek characters due to the
settings in the client machine.Maybe character data does not get converted, but still shown as greek.
Further, can any one tell us a way to verify what
characters get stored in the database and help us
identify if they are real Greek charactersIn the case of character data, what you really store depends on the db and the character values. You can use dump or some other PL function to look at the values.
There's a Database Globalization Support Guide for 9.2 that might be useful.