I'm trying to track down why appendObjectMatch (objectMatchTranslator) generates sql late in the process for a clause rather than immediately during objectNode processing.
So far, I don't understand. detectObjectMatch is called early in startNode to create/reset the objectMatchTranslator. appendLiteral and processRelTermination configure objectMatchTranslator during objectNode by adding DataObjects or Relationships or ObjectIds to the objectMatchTranslator. finishedChild (operator handler) stores sql and expression in objectMatchTranslator if matchingObject rather than to the global buffer. appendObjectMatch() is called after all of the objectNode and finishChild method calls during endNode.
