I agree, which is why I gave GROUP BY HAVING as an example filter for 'advanced' SQL skills (especially as my team manages data warehousing and analytics).
It's not a minor implementation detail, it's the major tool for filtering on aggregated values - something like 'find customers who have purchased two more products in one month' can either be done as a query of an aggregated sub-query, or as one query using HAVING to filter for customers with a COUNT(*) > 2.
Again, this is only for people claiming advanced skills, and it's not that an 'advanced' feature. I don't expect every developer to know about SQL:2003's window functions, but HAVING has been around since at least 1992.
I agree with not asking silly technicalities, as a Java developer, knowing why new Integer(100) == new Integer(100) but new Integer(1000) != new Integer(1000) is an interesting piece of trivia, but using it in an interview is a waste of everyone's time.
Syntax would be caring if they put the ORDER BY in front of the GROUP BY.
The technical interviewers must have taken it as a challenge to them, as they spent 30 minutes quizzing me on all of the different executables that made up the JDK and all of the different arguments to them.
This one. If you say you are expert/advanced at X, we will check for some level of competence. It's amazing how many candidates with CVs claiming 'advanced SQL skills' don't know about the existence of the HAVING clause. An exaggerated CV is a very easy filter.