Question

Consider this table ‘cat_post’:

entry_id    cat_id
2           2
2           4
2           76
3           2
3           4
3           76
4           2
4           76

I need to select the entries that have both a row with cat_id = 2 and one with cat_id = 4 (not all entries that have either 2 or 4), so this select:

select * from cat_post where cat_id IN (2,4)

will produce:

entry_id    cat_id
2           2
2           4
3           2
3           4
4           2

Which is not what I want. I need to get entries if they have both a row with 2 and one with 4, but not 2 or 4, like:

entry_id    cat_id
2           2
2           4
3           2
3           4

(without entry_id 4 in this case)

What query can I use to get this result?

Answer 1

You could use something like:

select entry_id from cat_post where cat_id in (2,4)
  group by entry_id
  having count(*) = 2;

This is extensible to more than two values in your “in” set, but will fail if (entry_id,cat_id) is not a unique key. If that’s not the case, use:

select entry_id from cat_post where cat_id in (2,4)
  group by entry_id
  having count(distinct cat_id) = 2;

Example:

mysql> select * from t;
+------+------+
| e    | c    |
+------+------+
|    2 |    2 |
|    2 |    4 |
|    3 |    2 |
|    3 |    4 |
|    4 |    2 |
|    5 |    4 |
+------+------+
6 rows in set (0.00 sec)

mysql> select e from t where c in (2,4) group by e having count(*) = 2;
+------+
| e    |
+------+
|    2 |
|    3 |
+------+
2 rows in set (0.00 sec)