This question has been prompted by the suggestion that this question is a dupe of this one. Substantively it is, but the solutions posted as answers are non-overlapping. Also in the first question, the test for do nothing is missing data, not a numerical logical test, and some solutions are specific to that case. They are not answers to the second question.
I feel that the newer question is more canonical and general and thus more likely to be found by future Googlers.
Is there an argument for defining duplicate narrowly and therefore not closing in cases like this?