Question

Assume the following tables:

TABLE: foo
=========================
| foo_id | fk_id | name |
=========================
| 1      | 100   | A    |
| 2      | 100   | B    |
| 3      | 200   | C    |
| 4      | 200   | D    |
| 5      |       | E    |
| 6      |       | F    |
-------------------------

TABLE: foo_combo
===============================================
| foo_combo_id | parent_foo_id | child_foo_id |
===============================================
| 1            | 5             | 1            |
| 2            | 5             | 2            |
| 3            | 6             | 3            |
| 4            | 6             | 4            |
-----------------------------------------------

I need to get all foo where fk_id is 100 and all foo combinations that are made up of foo where fk_id is 100. Using the sample data provided, I need fk_id = 1,2, (these have fk_id = 100) and 5 (this is a foo made up of foo that have fk_id = 100).

What should the SQL look like?

EDIT 1:

Two queries I think that need to be combined:

SELECT * FROM foo WHERE fk_id = 100
SELECT foo.* FROM foo, foo_combo WHERE foo_combo.child_id IN (SELECT foo_id FROM foo WHERE fk_id = 100) AND foo.foo_id = foo_combo.parent_foo_id ???

Answer 1

select foo_id from (
  select foo_id, fk_id from foo 
    union all
  select c.parent_foo_id, f.fk_id from foo_combo c inner join foo f on (c.child_foo_id = f.foo_id)
) foo_union where fk_id = 100;

That should return something like

+--------+
| foo_id |
+--------+
|      1 |
|      2 |
|      5 |
|      5 |
+--------+

The 5 is repeated because it matches twice. If you don’t want that, try adding “distinct” — ugly but serviceable.