Effective in version 10.5, the Spark engine increased performance of the Router transformation. When the Spark engine runs a mapping that includes a Router transformation, the Spark engine processes the upstream mapping pipeline once and stages the data so that it can be used by each downstream branch.

Previously, the Spark engine processed the upstream mapping pipeline separately for each output group.

For more information, see the "Router Transformation" chapter in the

Informatica 10.5 Developer Transformation Guide

.