Archives of the TeradataForum

Message Posted: Fri, 16 Apr 2004 @ 18:34:02 GMT

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

Thanks for clearing that up Rohit.

However, it does make the solution appear to be one of needing more SPOOL. It sounds as though when you were using a single table that the joined rows were all in a single SPOOL. Now, some are in one spool area and the other have their own as well. During the step that fails, Teradata may be placing copies of the two spool files into a third and not finding enough allowed. Looking at an explain is the only way to know for sure.

If you are doing this separation for speed, than you have come across a great reason for using PPI (Partitioned Primary Index). This allows for Teradata to scan only a fraction of the rows (one or more partitions) of a large table instead of scanning all of the rows. This is a feature in V2R5.

I could make many of your queries faster, including this one and cut down your SPOOL utilization.

Michael Larkins
Certified Teradata Master

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