Archives of the TeradataForum
Message Posted: Wed, 09 May 2001 @ 11:10:11 GMT
| Subj: || || Re: Unique Row ID |
| From: || || Claybourne Barrineau |
Just out of curiousity, are you performing a csum(1,1) to get a unique row id. I tried this once, and I got the same results as you (hot
amp.) However, if you perform a csum(1,('Primary Index Component #1' || 'Primary Index Component #2' || .....)) (assuming the PI is unique
and you can maintain uniqueness somehow with a concatentation and all concatenated columns are not null), I think this will help. In
theory, this approach should distribute your csum() across the amps based upon the distribution of your PI.