This implies ditching the JcrConfigurationSource.RawViewProducingNode2BeanTransformer. There are several reasons for that:

• it is complicated to both use the ToBeanTransformer API provided by N2B processor and pass a custom N2B transformer to it (helper API will be present for that in the mean time)
• if N2B transformation fails for some reason - the raw view might be incomplete
• It is complex - there're several workarounds and overall some fuzzy logic in that class. It would be much simpler to walk the node tree separately from N2B transformation and compute the raw view directly from the JCR data (some possible concerns though might include type clashes)