empty struct fields through an adapter (here Bridge Adapter

I send documents from an ATC (V4.1) to an Integration Server (B2B V4.01), through an Enterprise Broker (V4.1.0) and then a Bridge Adapter (V2.0).
These documents contain some empty fields and empty struct fields.
How does “RecordToDocument” (in Integration Server) handle such fields to generate resulting XML file?
Behaviours concerning such empty fields are not documented, and are unexpected for me:

1 - The empty struct fields or empty fields are also generated in XML. This implies later treatements in the Integration Server Side before sending the I.S. documents to the target application.

2 - When “allow unset fields” is unchecked in E.I. “character 0” fills the empty fields in Integer XML output.

3 - When “allow unset fields” is checked in E.I., no character is inserted but the empty fields are still present in XML output.

Are these behaviours normal behaviours?
Is there a way to suppress these empty fields from the XML output?

This is expected behaviour of the Bridge, i have a small java utility to solve this issue. let me know if you are still interested.