Question

I want to get row count in tableA if and only if rowA_x does not have a FK pointing to rowB_x in tableB.

tableA:

id | id_tableB

tableB

id | ...

So basically the rows in tableA should only be counted if the column id_tableA does not exist as id in tableB.

Is there a clean way to do such counting. I have around ~500.000 rows.

Answer 1

There are several (not 100% sure about the MySQL syntax, so this might require some tweaking):

Subselect with NOT IN:

select count(*) from tableA where id_tableB not in (select id from tableB);

Subselect with NOT EXISTS:

select * from tableA a 
where NOT EXISTS (select null from tableB b where a.id_tableB = b.id);

OUTER JOIN:

 select count(*) from (
    select a.*, b.id as b_id
    from tableA a
    left join tableB b on a.id_tableB = b.id)
  where b_id IS NULL;

Which of these is the fastest depends on your data, but usually, a JOIN is more efficient than a subquery.