Question

I’m having problems selecting the right data from a table with one query. I’m
unsure if it can be done.

properties:
- id
- name

facilities:
- id
- name

property_facilities:
- property_id
- facility_id

As you can see in the schema above, each property can have one or more
facilities. At a certain point in my code I need to have a list of property ids
that have some facilities.

PSEUDO SQL:

SELECT property_id
FROM property_facilities
WHERE
  facility_id IN (8, 12)

Obviously this doesn’t work, since I need to have a list of property ids that
have BOTH the facility 8 and 12, not just one of them. How could I achieve
this?

Answer 1

Adding a HAVING clause with the expected number of matching id’s will do it:

SELECT property_id
FROM property_facilities
WHERE facility_id IN (8, 12)
GROUP BY property_id
HAVING count(facility_id) = 2

This does assume, however, that you have no duplicate entries in your property_facilities table. If you do, you can fix this using DISTINCT in a subquery, or by simply adding constraints not to allow duplicates in the first place.

Here’s the same query with a DISTINCT subquery:

SELECT property_id
FROM (SELECT DISTINCT property_id, facility_id FROM property_facilities) distinct_rows
WHERE facility_id IN (8, 12)
GROUP BY property_id
HAVING count(facility_id) = 2