Archives of the TeradataForum

Message Posted: Tue, 28 Oct 2008 @ 22:25:50 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hi Phil:

How randomized do you need the sample?

If it does have to be extremely random, you could do the following:

     qualify count(*) over (partition by location
                             rows unbounded preceding) <6

Hope this helps,

Michael Larkins
Certified Teradata Master
Certified Teradata SQL Instructor

	<Prev	Next>		<<First	<Prev	Next>	Last>>