|
|
Archives of the TeradataForum
Message Posted: Wed, 24 May 2000 @ 15:05:19 GMT
Subj: | | Re: Surrogate keys |
|
From: | | Jim Downey |
I don't know if this will help but we use a process where we compare the new data to the old data. If the new data is represented in the
old data, we inherit the surrogate key. If the new data is not represented (and it needs a new key) then we rank the data using the RANK
function. This gives each row a number from 1 to n. We then take max(old data) and add it to the new data. For example if our old data
had keys from 1 to 1000 and the new data had ranked from 1 to 20, when we merge the new data with the old then we add the max (1000) to each
rank giving us new values from 1001 to 1020.
| |