Question

I have following query:

select id from t1
intersect
select id from t2
intersect
select id from t3

id could be non unique in some tables, so I need to use distinct.

In general what is better:

select distinct id from (
select id from t1
intersect
select id from t2
intersect
select id from t3)

or

select distinct id from t1
intersect
select id from t2 -- here id is unique
intersect
select distinct id from t3

Answer 1

There is no need for the DISTINCT. The INTERSECT operator automatically produces a distinct set of values. As you can see in this example, x has two rows with an ID of 1, y only has one row with an ID of 1 and the INTERSECTION of the two produces just a single row

SQL> ed
Wrote file afiedt.buf

  1  with x as (select 1 id from dual union all select 1 from dual),
  2       y as (select 1 id from dual)
  3  select id
  4    from x
  5  intersect
  6  select id
  7*   from y
SQL> /

        ID
----------
         1

Even if you take the INTERSECT of the two row table with itself, you still get a single row in the output

SQL> ed
Wrote file afiedt.buf

  1  with x as (select 1 id from dual union all select 1 from dual)
  2  select id
  3    from x
  4  intersect
  5  select id
  6*   from x
SQL> /

        ID
----------
         1