Aggregator transform is m uch like the Group by clause in traditional SQL.
this particular transform is a connected/active transform which can take the incoming data form the mapping pipeline and group them based on the group by ports specified and can caculated aggregate funtions like ( avg, sum, count, stddev....e.tc) for each of those groups.
From a performanace perspective if your mapping has an AGGREGATOR transform use filters and sorters very early in the pipeline if there is any need for them.
veepee
