 |
 |
Archives of the TeradataForum
Message Posted: Wed, 10 Dec 2008 @ 14:39:52 GMT
| Subj: | | Re: Rank and Delete? |
| |
| From: | | Victor Sokovin |
| | Good news and bad news. That worked great in SQL Assistant, but resulted in an invalid statement error (3537: A MultiLoad DELETE Statement
is Invalid.) when trying to run it in an MLOAD job. Apparently MLOAD deletes can not use subselects and must key on the primary index. At least
that is what I am being told by other folks here. | |
| | Oh well. I guess I am back to a multi-step process; an EXPORT followed by the MLOAD Delete. | |
Bob, your original posting does not seem to mention the intended use of the statement in ML so I would not classify your news as good or bad,
I'd merely say that this is news for us.
Teradata bulk utilities are intended for fast processing of large volumes of data but they come with restrictions on the SQL they can take care
of. It is not as simple as taking any query out there, submitting it via the bulk utility, and hey presto.
I'd probably second Michael and recommend not to use ML if you are not comfortable with it as yet. If for some reason you really want to use it
in this particular case then I would recommend reading the ML manual before proceeding with the task. The manual provides information on
differences in DELETE statements in the so called "Import Tasks" and "Delete Tasks". The latter are less restrictive in their syntax, for example,
but they would imply a full table scan and you are probably not much better off performance-wise than with BTEQ.
So, read the manual, try out a few scenarios using different utilities, measure the run times and decide what works best for you.
Victor
| |
