UNICODE CSV files
The ODBC data source administrator tool has a couple of limitations when it comes to supporting UNICODE. The editor does not properly detect/handle the byte order mark (BOM) that is present in many UNICODE files. Nor is UNICODE an option for the character set in the editor.
To support UNICODE CSV files, you need to manually override the settings in the schema.ini file.
For UTF-8 files you must:
Manually change the CharacterSet= key-value in the schema.ini to:
CharacterSet=65001Manually define the Columns' names and data types since the Guess option does not properly handle the BOM.