MySQL count all records that have at least one record on a pivot table

Question

I'm trying to look if a record of a table, has at least one record on a pivot table that is used on a many to many relationship.

The table Attendant is:

| ID |  NAME  |
|----|--------|
| 1  | Name A |
| 2  | Name B |
| 3  | Name C |

And the Attendant_Event pivot table has the following structure

| ID |  attendant_id  |  event_id  |  uuid  |
|----|----------------|------------|--------|
| 1  |       1        |      1     |   xxx  |
| 2  |       1        |      2     |   yyy  |
| 3  |       3        |      1     |   zzz  |
| 4  |       3        |      2     |   www  |
| 5  |       1        |      3     |   xyx  |
| 6  |       3        |      3     |   rer  |

My query is trying to count the attendants that has a least one record on the pivot table, but count all the records as one. For example, the expected result will be a table like this:

| STATUS |  COUNT |
|--------|--------|
|   YES  |    2   | 
|   NO   |    1   |

This results are expected because:

1. Only the Attendants with ID 1 and 3 has a record on the Attendant_Event table. This told us that the number of Attendants with rows on the pivot table is 2.
2. The Attendant with ID 2 has no records, so the number of attendants with no records on the pivot table is 1.

By now, my query is the following:

SELECT IF(uuid <=> NULL, 'NO', 'YES') as status, count(*) as count FROM attendants att LEFT JOIN attendant_event ae ON ae.attendant_id = att.id GROUP BY status

But this is showing me a result like this.

| STATUS |  COUNT |
|--------|--------|
|   YES  |    6   | 
|   NO   |    1   |

This means that, count each of the rows. If we take the previous example, both Attendants with id 1 and 3 has 3 records on the pivot table. So It gives 6 instead of the two that I'm looking for.

What I'm doing wrong?

Answer 1

You may want to select the attendant IDs with their respective YES/NO first, then count them, something like:

SELECT status, count(distinct attendant_id) as count FROM (
   SELECT IF(ae.uuid IS NULL, 'NO', 'YES') as status, ae.attendant_id
   FROM attendants att LEFT JOIN attendant_event ae ON ae.attendant_id = att.id
   GROUP BY ae.attendant_id) x
GROUP BY status

Answer 2

When you make left join, you create intersection, that is larger than attendants table. Your join consists of rows with repeating attendant_id and different event uuid .

You can watch the intersection by executing SELECT IF(uuid <=> NULL, 'NO', 'YES') as status, att.id, ae.uuid FROM attendants att LEFT JOIN attendant_event ae ON ae.attendant_id = att.id . It includes 7 rows, 6 of them are with YES events for two active attendats and 1 row with NO events.

So you should count only distinct values:

SELECT IF(uuid <=> NULL, 'NO', 'YES') as status, count(distinct(att.id)) as count 
   FROM attendants att 
   LEFT JOIN attendant_event ae ON ae.attendant_id = att.id 
   GROUP BY status