 |
 |
Archives of the TeradataForum
Message Posted: Thu, 25 May 2000 @ 18:03:05 GMT
| Subj: | | Re: Surrogate keys |
| |
| From: | | Bethany Withers |
Our mature (aka old) applications all use intelligent keys, for example store, item. We have a new application in very early design
phases that will be using surrogate keys. It deploys early 2001, so we will have some experiences to share in 8-9 months. We have had huge
success by using multi-set tables for our large historical tables. By dropping date out of the key our I/O dropped dramatically and our
online queries run significantly faster. Batch load time increased slightly. You may want to test multi-set with 64K block size (available
in V2R3) for your table where policy is not unique enough.
| |
