I am facing a similar issue while processing arabic text. In fact I have to process text files containing arabic text. The default file encoding is ASCII, I convert it into UTF8, that is fine. The data is loaded, but I can see only garbage. If anyone hav faced this issue, please guie!
If I remember correctly, there are ODBC properties and change the encoding/character set and you can view the resultset in SQL Assistant. Its only the display problem becasue the SQLA is not able to show it properly, the data loaded should be fine.
I will try to see what exactly was changed to see the arabic data, and will let you know if I am able to find it :)
You can verify your loaded data using CLI based connectivity i.e. BTEQ.
Also, about the setting that Qaisar has mentioned .... you need to go to ODBC -> Edit TD connection -> Change 'Session Character Set' from ASCII to UTF8.