|
|
Archives of the TeradataForum
Message Posted: Thu, 18 Oct 2007 @ 10:06:50 GMT
Subj: | | Re: Removing duplicates from table |
|
From: | | Curley, David |
To follow on what Frank said, "first" doesn't mean anything in RDBMS, not just parallel systems.
If you're doing it in one shot, decide what makes something "first" and qualify on row_number() over (partition by UPI order by "first
criteria") = 1.
If you'll be loading more than once or from more than one source table, you'll need to do that and where not exists (select 0 from target_table
where target.upi = source.upi).
Dave
| |