I cannot give you the "reverse of the same problem" because I still do not understand the original problem. I stated that the code I posted did
not pass your single test case. I posted it to demonstrate my understanding of the problem. I expected you to post a clarification.
You now mention "Partially mismatched" lines. I cannot think of any interpretation of this phrase which is consistent with your new test case. In addition to all the suggestion from AnomolousMonk, I also tried: "Select a line from file1 if it does (not) contain any word which appears in file2" (Where "word" is defined as all the text between regex word boundaries.)
Please post unambiguous requirements and several test cases. It is important that they all be exactly correct.