|
|
Archives of the TeradataForum
Message Posted: Tue, 19 Feb 2008 @ 21:02:50 GMT
Subj: | | Re: SQL to populate additional data |
|
From: | | Curley, David |
Hi Michael -
Michele doesn't want to add 9 rows, only 8. There's already a row with product/expanded_product 0111/0111, so there's no need for
011X/0111.
But we weren't told that for _every_ product, only the equivalent to the 011X/0110 and 0111/0111 rows are present. What if the next product
has 012X/0121 and 0125/0125? Or what if there are already a couple X rows, or more than one non-X row for the same 3 leading characters? Since
we don't have the full picture, I'd lean towards the more generic approach I outlined earlier.
If we know that we only have ###X/###0 and ###1/###1 rows, the union approach would totally work for numbers 2-9, although I'd worry about
overflowing a byteint since we don't know how high they go (we do know there are 2,000 rows, so some numbers have to be > 128).
Dave
| |