|
|
Archives of the TeradataForum
Message Posted: Tue, 30 Oct 2007 @ 12:00:01 GMT
Subj: | | Re: Generate unique id |
|
From: | | Dieter Noeth |
Barner Eric wrote:
| This is true, also I think it inherently has higher tendency to skew more so than the ROW_Number function. | |
| Or the SUM(x) over(partition by ..... Order by ... Rows unbounded preceding) form. | |
| Can anyone from Teradata confirm the tendency to skew with CSUM? | |
I'm not from Teradata, but AFAIK CSUM is exactly the same as SUM OVER, it's skewing if there are not enough distinct PARTITION BY/ORDER BY
values. Most stupid example would be CSUM(1,1).
Dieter
| |