Question

Is there a better way to perform this query without using a nested subquery?

-- select all races for jockeys which have multiple regions
select distinct r.id, r.description
from jockeys_races jra inner join races r on r.id = jra.race_id
where jra.jockey_id in (
    select jre.jockey_id
    from jockeys_regions jre
    group by jre.jockey_id
    having count(*) > 1
);

Jockeys can belong to multiple regions. Jockeys can be in multiple races.

Answer 1

It may not be as efficient (since the product will be larger), but you can do it. Check the EXPLAIN on both.

select distinct r.id, r.description
from races r
  inner join jockey_races jra on r.id = jra.race_id
  inner join jockey_regions jre on jra.jockey_id = jre.jockey_id
group by r.id, r.description, jre.jockey_id
having count(*) > 1