Archives of the TeradataForum

Message Posted: Tue, 28 Oct 2003 @ 18:56:14 GMT

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

Narayan,

Henry is right that usually you won't see a group amp step unless 50% or fewer amps are involved in the query. But the case you asked about (shown below) looks like a slightly different use of group amp:

The result goes into Spool 1 (group_amps), which is built locally on the AMPs.

This looks like it could be an extract from a final all AMPs retrieve step. In V2R5 when a query places final results into the last spool (usually Spool 1) the DBMS tries to limit the amps involved to just those amps that actually have rows. Previously, all amps in an all amps retrieve step were always involved in final response processing. This group amp operation optimizes response processing by only including amps that actually have rows to be returned in the BYNET merge activity. In some cases this may be all the amps, in other cases 80% of the amps, and in some cases 1 amp.

Here's an example of the final step from an all amps query:

7) We do an all-AMPs RETRIEVE step from Spool 5 (Last Use) by way of an all-rows scan into Spool 1 (group_amps), which is built locally on the AMPs. The size of Spool 1 is estimated with high confidence to be 1 row. The estimated time for this step is 0.02 seconds.

Thanks, --Carrie

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