Archives of the TeradataForum
Message Posted: Thu, 12 Apr 2007 @ 11:46:10 GMT
| Subj: | | Re: Merging large volumes of data into large history tables |
| |
| From: | | Walter, Todd A |
Of course nothing is ever fast enough but...
One has to consider the full end to end time and/or being constrained from utilizing features. The reason for moving these operations
increasingly to the SQL engine rather than a specialized utility is to remove limitations such as Multiload not allowing Join Index. If you have
to DROP JI, Mload and CREATE JI, it is very likely that the SQL operation that maintains the JI in place will be much better than the drop create
process.
Plus I think that those of you who have not performance tested the SQL DML operations in a while will be very pleasantly surprised when you do
so on V2R6.1, 6.2 and TD12.0.
| 
