Splitting large documents using a Data Process shape may result in data loss (see community post).
A workaround to this issue is to split the document using a map. In the map, the maxOccurs of the Destination element in the source profile is set to unbounded. In the target profile, the maxOccurs of the element Destination is set to 1 (one).
The map contains the map function Set DDP_ROUTE_DESTINATION. This function trims, converts to upper and results the output and placing it in the dynamic document property DDP_ROUTE_DESTINATION. The goal is to extract the routing data for each document and split the document at the same time, minimizing the need to access the document data.
The map splits the test message correctly into 6 documents, each containing a single Destination element with a correct and expected value (A4F, MAG, SF, SHP, SB and AAA) in the sub element EndPointCode. The unexpected behavior is that the dynamic document property DDP_ROUTE_DESTINATION has the value AAA on all split documents.
Is this a bug or by design?