Question

I have a table, which looks like:

+-----------+----------+
+ person_id + group_id +
+-----------+----------+
+    1      +    10    +
+    1      +    20    +
+    1      +    30    +
+    2      +    10    +
+    2      +    20    +
+    3      +    10    +
+-----------+----------+

I need a query such that only person_ids with groups 10 AND 20 AND 30 are returned (only person_id: 1). I am not sure how to do this, as from what I can see it would require me to group the rows by person_id and then select the rows which contain all group_ids.

I’m looking for something which will preserve the use of keys without resorting to string operations on group_concat() or such.

Answer 1

This is what you’re looking for:

select person_id from t
where group_id in (10, 20, 30)
group by person_id
having count(distinct group_id) = 3

Although efficient, using this solution the amount of values in the in will have to match the value you compare the count with.

And the just-for-fun solution, as you said, even with group_concat you can solve this 😛

select person_id from t
where group_id in (10, 20, 30)
group by person_id
having group_concat(distinct group_id order by group_id) = '10,20,30'