1 2 3 4 5
--!syntax_pg select x,count(y) from (values (1,2),(3,4),(3,5)) u(x,y) group by x having count(y)=2