 |
 |
Archives of the TeradataForum
Message Posted: Mon, 25 Feb 2008 @ 21:21:47 GMT
| Subj: | | Re: TPump questions |
| |
| From: | | Michael Larkins |
Hello David:
TPUMP can be a bit tricky to setup and run due to its row at a time nature. To help you as much as I can:
input record used in INSERT/UPDATE would create a new row that is a duplicate of an existing row - IGNORE = don't put in to error table,
default is MARK = put into error table
input record used in DELETE/UPDATE does not match a row in a table
Because Teradata supports NUPI, if an input record to an UPDATE/DELETE causes an activity count > 1, extra rows processed
| | We're using an Upsert an I'm not sure what these options can do for me. | |
UPSERT (UPDATE/INSERT) to same table in same DML LABEL requires the DO INSERT FOR MISSING UPDATE ROW error treatment option on >DML LABEL
command. When this is specified, IGNORE is the default for MISSING UPDATE ROW
| | Also, if we've got a Multiset Table with only a NUPI on the table, so whilst the Upsert will mention the full logical primary key,
does the table actually have to have a unique constraint added to the table ? | |
Not really, but it might be a good idea - see EXTRA ROWS above
| | Lastly, I've noticed that a Macro is created by the utility, | |
My recommendation is to ALWAYS run TPump and leave the macros alone. Otherwise you are over-riding all TPump error handling, recovery, input
processing, and the works.
Regards,
Michael Larkins
Certified Teradata Master
Certified Teradata SQL Instructor
| |
