This is a common query generated by Hibernate

SELECT b.*
FROM b WHERE b.a_id IN
(SELECT a.id
FROM a
WHERE a.id IN (100, 200))

If selecting the same column with the same value, this could be constant-folded as (100, 200)
Otherwise, this involves opening two tables and doing index lookups. Since this is such a common pattern 50% of all our hibernate created queries, it makes sense to try filing a bug for it.

More generally this optimization can be for any operator against any key

MySQL cannot optimize it because it believes there could be rows in b with a_ids not present in a. But since we enforce this in Hibernate (not as a foreign key) it makes sense for hibernate to be smart her.