Web connector vs HTTP connector

@Mark Jongeling the scenario.  Customer integration with Klippa. 

In the demo we used the HTTP connector.  Where I generate the request body from a documents table with documents I need to scan. 

select crd.document_data as [data]
      ,crd.document_upload as [filename]
      ,crd.document_page_ranges as [page_ranges]
  from credit_request_document crd
 where crd.credit_request_id          = @credit_request_id 
   and crd.credit_request_document_id = @credit_request_document_id
   for json path, root('documents')

With the HTTP connector the result of the query is the request_body towards the klippa API. 

Now I want to have this API as a web connector so I can differentiate API keys per environment. 

I put everything in place and we end up setting up the body.  Type JSON. 

option 1) I put variable in Body {request_body} results in non parsable body (whilst with HTTP connector this works.  There I also put the variable as body. 

option 2) I put in the json structure and fill every element with data from the table. Fails with unparsable content. I tried all options to also have input parameters pre processed.. (although I find it very unclear what it does exactly) 

{ documents: [
  { data : {document_data},
    filename : {document_upload},
    page_ranges : {document_page_ranges}
   )
  ]
} 

option 3) this one finally worked..  I put the body request JSON root element and just made the array a variable.  I really have not idea why only this one worked and did not result in a non-parsable document.  

{
    "documents": {document_array}
}

When I read the documentation, I can read it, but don't understand it.  On the pre-processing part: "so that Indicium can use it"…  ???  It really doesn't help me to understand when to use it or not to use it.  

https://docs.thinkwisesoftware.com/docs/sf/web_connections#web-connection-for-a-branch