Yeah, but any competent DB admin will at least read the logs, do a few A/B tests, blocking out different parts or using queries which should be equivalent and then say 'something is hinky with the way this query is compiling, query z seems to give a different answer if query y is written like foo even though I see no link. Probably a subtle bug in your code I haven't spotted. Maybe ask someone who knows more about the compilation step?'
Even if they 'solve' it by rewriting query y, they'll not claim that query y was the problem, and may even leave a comment or commit log saying they didn't understand why writing it like har fixed it.
Even if they 'solve' it by rewriting query y, they'll not claim that query y was the problem, and may even leave a comment or commit log saying they didn't understand why writing it like har fixed it.