Archives of the TeradataForum
Message Posted: Thu, 02 Feb 2006 @ 10:16:43 GMT
| Subj: | | Re: Use of Multiset for Large Table |
| |
| From: | | Anomy Anom |
<-- Anonymously Posted: Thursday, February 02, 2006 04:51 -->
| | I think this difference is seen only with insert-select vs normal insert. for a normal insert both set and multiset errors out, whereas
for insert-select it errors out only for multiset table. For set table, the duplicates are just ignored and no errors are generated. | |
You are right. But it is rationally inexplicable why it should behave this way. Also,I think allowing UPI on multisets is again from absurd
from a logical perspective.
I have also observed fastload loads data faster for a multiset table, though it is designed not to load duplicates. Here again, no logical
reason why allow multiset tables with fastload.
That we all exploit multisets to achieve performance is no question. But some of the above are to me antithetical to logic.
(Teradata aint mathematics to give tautological truths!!). Are we at the mercy of the whims of the product designer??
Cheers
| 
