Seems like it could mask all sorts of errors. The query could be correct, but the data model could have been setup incorrectly allowing duplicate data to get into the system. Or a uniqueness constraint could be dropped by accident and go undiscovered for a very long time. This is similar to assuming a catch placed around your code is going to help you find a bug.
My purpose in posting the above quote was in hopes that someone with real world experience would explain what was wrong with this reasoning. Thanks Don!
