I got 4/5 on PostgreSQL. I selected "Not enough information" on question 5, which is supposed to be the wrong answer. But I think it's the correct answer.

Depending on the distribution of the data in the table (e.g. all rows have a = 38, but no rows have b = 1), PostgreSQL will chose a Sequential-Scan instead of an Index-Only scan for both queries. This of course leads to the 2nd query being faster due to not having to aggregate anything.

https://gist.github.com/felixge/af97c844cb1f24ac278e0357741c...