Prior to the tXMLMap, you can use a tMap and tFilterRow to establish a reject output.  Left outer join the lookup to the DB customers, passing along the db's id which will be null if the lookup fails.  tFilterRow records with a db id.  The filter connection goes on to tXMLMap while the reject connection goes to a second tXMLMap which produces XML specific to the reject handling.

The reject handling could be a separate XML document.  Or, you can use 2 tAdvancedFileOutputXMLs in append mode instead of the tXMLMap.  See the section "Multiple Loops" in the blog post http://bekwam.blogspot.com/2011/09/xml- … urces.html for how to do thi.

Sorry for a delay, I did not get a notification the response was available.
Can you clarify please how to use tFilterRow ? I've uploaded two more images, tFilterRow1.png and tFilterRow2.png.

tFilterRow1.png shows a tFilterRow being inserted between tRESTRequest and tXMLMap, with Customers (DB) acting as a lookup. Note that 'id' which is coming from tRESTRequest is never null or empty, so the tFilterRow/Reject is never activated, if no record exist I simply get an empty customer XML back,

tFilterRow2.png shows a tFilterRow being inserted between Customers and tXMLMap, in the actual job it has a 'NAME NotEqualTo ""' rule (CUSTOMERS return ID & NAME).
When I have an existing record being retrieved, tFilterRow directs it to the main output which is fine, and if I request a customer with a non-existent id, the job hangs and the tFilterRow rejects output is not activated.

Sorry if I'm missing something

Thanks a mill, adding an extra tMAP made it work, see the attached image.