sqlalchemy and double outerjoin

Question

I need to do double outer join on following 2 tables A and B to get presented result using SQLAlchemy ORM or SQL expressions.

Table B should be outer joined twice to get joined 2 result sets (distinguished by c_id) that are for the same A records. Outer join is used to get NULLs where B results are missing either in first (c_id = 66) or second (c_id = 70) outer join.

A table:

id
--
1
2
3
4

B table:

id | a_id | c_id
---+------+------
1  |  1   | 66
2  |  2   | 66
3  |  3   | 70
4  |  4   | 66
5  |  4   | 70

Query result should be:

a_id | b1_id (66) | b2_id (70)
-----+------------+-----------
 1   |  1         |  NULL
 2   |  2         |  NULL
 3   |  NULL      |  3
 4   |  4         |  5

I got to the point where proper raw SQL query looks as follows:

SELECT 
  A.id AS a_id,
  B_1.id AS b1_id, 
  B_2.id AS b2_id,  
FROM 
  A
LEFT OUTER JOIN   B AS B_1  ON A.id = B_1.a_id AND B_1.c_id = 66
LEFT OUTER JOIN   B AS B_2  ON A.id = B_2.a_id AND B_2.c_id = 70
WHERE
  B_1.id is not NULL or 
  B_2.id is not NULL;

Now, do you know how to get this coded either in SA ORM or SA SQL expressions?

Answer 1

Found the solution myself:

b1 = aliased(B)
b2 = aliased(B)
q = session.query(A.id, b1.id.label("b1_id"), b1.id.label("b2_id"))
q = q.outerjoin(b1, sqlalchemy.and_(A.id == b1.a_id, b1.c_id == 66))
q = q.outerjoin(b2, sqlalchemy.and_(A.id == b2.a_id, b2.c_id == 70))
q = q.filter(sqlalchemy.or_(b1.id != None, b2.id != None))