In a Pass Through session, Unicode characters (including PTCs) are imported to a UNICODE server character set as follows:
- Teradata supported characters in 6.0 BMP are processed as 2-byte UTF16.
- Teradata unsupported characters in 6.1 to 9.0 BMP and all SMP are converted to and processed as 2 or 4-byte UTF16.
- Unassigned and Private Use characters in the BMP and SMP are converted to and processed as 2 or 4-byte UTF16.
- Noncharacters are changed to the REPLACEMENT CHARACTER (U+FFFD).
- Ill-formed code unit sequences are changed to the REPLACEMENT CHARACTER (U+FFFD).
- CLOBs are supported.
In sessions where UPT is not enabled, importing a PTC results in an External-to-Internal (E2I) conversion exception and the query fails.
UPT only applies to Unicode data and sessions, meaning UTF8 and UTF16 sessions and the UNICODE server character set. For example, importing the 0x1A replacement character from an ASCII character set to a Unicode column will still be rejected even in a session enabled with the UPT feature.
Users should also be aware that once PTCs are imported into the database, they can percolate through the system so there is a possibility that sessions with UPT disabled may still encounter PTCs. If you want to remove PTCs from your system, you must delete or replace the PTCs.