For example, create two indexes on both obj1 and obj2
create index a_index on bucket1(obj1,status) using gsi;
create index b_index on bucket1(obj2,status) using gsi;
when run this query, it only use index scan on “a” using a_index. Is it using key scan on “b”? Is there a way to involve b_index?
select *
from bucket1 a join bucket1 b keys a.type
where a.obj1.status=“true” and b.obj2.status=“true”