Archives of the TeradataForum

Message Posted: Wed, 09 May 2001 @ 19:28:27 GMT

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

Just to add my two-penny'th to this ...

Using CSUM(1,column-name) will always generate unique values, even if 'column-name' contains non-unique values. Try it on a small data set and see.

Also, the more unique the values for 'column-name' (and yes you can use multiple columns) then the more even will be the distribution of the spool file.

Cheers,

Dave

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