Although UPT allows the storage and retrieval of PTCs, full support is not available for these characters in Teradata. For example, the following are not supported for PTCs:
- Collation
- Case sensitivity
Uppercase function, uppercase storage, and lowercase function may produce incorrect results.
- Object name support
PTCs cannot be used as part of an object name.
- Inequality comparisons may produce incorrect results.
- PTCs should not be stored in a column that is part of an index.
Because casing, collation, and object support are not included with UPT, there are some scenarios where the behavior will differ from what is expected with fully supported characters. For example, the following may produce unexpected results:
- Value ordered NUSIs or join index with ORDER BY clauses
- RANGE PPI definitions
- Case blind equality comparison and uniqueness constraints
- Upper and lower case translation
- Interaction with PIVOT (PTCs cannot be column names)
UPT does not resolve these conversion errors:
- NonUnicode client to Unicode (on error character)
- Extended site-defined client character set (for example, ARABIC1256_6A0) to the Latin server character set
- Unicode to Latin
- Latin to Unicode (error on 1A)
- TRANSLATE without WITH ERROR with nonUnicode
- KanjiSJIS to Unicode or Latin (error on 1A)
The following additional restrictions and limitations apply:
- UPT does not provide a per column granularity.
- Standalone FastExport, FastLoad, MultiLoad and Tpump do not support PTCs. You can use Teradata Parallel Transporter to load and unload data containing PTCs.