When you configure relational output, you can configure a separate output group for each multiple-occurring input node. You can also create groups that contain denormalized data. You can pivot multiple-occurring elements and limit the number of occurrences in an output group.