Restrictions and Limitations
This section describes the restrictions and limitations when using the Update operator.
Table Characteristics and Considerations
When using the traditional MultiLoad protocol, the target table cannot have any of
the following characteristics:
Join, hash, or secondary indexes
No primary index
Foreign key references
Referential integrity
Triggers
Cannot be a normalized table
Cannot be a column partitioned table
Cannot be a temporal table
Cannot be in a replication group
The extended MultiLoad protocol can be used to load target tables that cannot be loaded
using the traditional MultiLoad protocol. The exceptions are:
Cannot be a column partitioned table
Cannot be a temporal table
Cannot be in a replication group
Data Types
The job cannot load these data types:
LOB
JSON
XML
ST_Geometry
UDT data can be loaded in its external type format.
Error and Work Tables
The table names for the ErrorTable1, ErrorTable2, and WorkTable attributes must be
new tables unless you are restarting a paused Update operator job.
Other Notes
One Update operator job can load up to 5 Teradata tables for an import task. A single
table is supported for a delete task.
By default, duplicate rows are not discarded by the Teradata Database for SET tables.
Duplicate rows are inserted into the application phase error table. Use the DML option,
"IGNORE DUPLICATE ROWS," to override the default if you do not want the duplicate
rows to be inserted into the application phase error table.
Duplicate rows are not discarded by the Teradata Database for MULTISET tables; they
are inserted into the target tables. The DML options, "MARK DUPLICATE ROWS" and "IGNORE
DUPLICATE ROWS" cannot override this behavior. Duplicate rows are allowed for MULTISET
tables.
Only the Teradata SQL INSERT, UPDATE, DELETE, and UPSERT statements are supported;
any other Teradata SQL statements are not supported.
Data can be loaded into empty or populated Teradata tables.
A delete task cannot be used on a view.
The target tables are locked until the application phase is complete.
The target tables cannot have column names that are the same as the error table column
names, because the application phase error table includes a mirror image of the target
table, preceded by the error information. For more information on the application
error table "Update Operator Errors" in Teradata Parallel Transporter User Guide.
You can pause the job for an import task. You cannot pause the job for a delete task.
The target tables cannot be rolled back after the application phase is complete.
The Update operator requires a Teradata Database load slot.