Archives of the TeradataForum

Message Posted: Mon, 04 Apr 2005 @ 15:23:36 GMT

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

Mirjam:

Spool 2 is the input to the stat fn step

Spool 9 is the statfn work spool (partitioned and ordered to meet the ordered analytics processing requirements)

Spool 1 is the output spool

There are two partitioning mechanism that a stat functions work spool can use; VALUE based or HASH based. The value based mechanism is intended to enable parallelism for the cases where there are a large number of rows and a few number of groups. The hash mechanism is used for all other cases of rows/groups. The partitioning mechanism is chosen at runtime based on sampling the input data and therefore the compile time explain text is only an assumption of the actual partitioning.

Mike

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