Turn a Mysql Subquery in a Join

Question

How can I turn this subquery in a JOIN ?

I read that subqueries are slower than JOIN s.

SELECT 
    reklamation.id, 
    reklamation.titel,
    ( 
        SELECT reklamation_status.status 
        FROM reklamation_status 
        WHERE reklamation_status.id_reklamation = reklamation.id 
        ORDER BY reklamation_status.id DESC 
        LIMIT 1 
    ) as status
FROM reklamation 
WHERE reklamation.aktiv=1

Answer 1

Using JOIN query can be rewritten to:

SELECT reklamation.id, reklamation.titel, reklamation_status.status
FROM reklamation 
JOIN reklamation_status ON reklamation_status.id_reklamation = reklamation.id
WHERE reklamation.aktiv=1

Answer 2

This should do it:

SELECT r.id, r.titel, MAX(s.id) as status
FROM reklamation r
LEFT JOIN reklamation_status s ON s.id_reklamation = r.id
WHERE r.aktiv = 1   
GROUP BY r.id, r.titel

The key point here is to use aggregation to manage the cardinality between reklamation and reklamation_status . In your original code, the inline subquery uses ORDER BY reklamation_status.id DESC LIMIT 1 to return the highest id in reklamation_status that corresponds to the current reklamation . Without aggregation, we would probably get multiple records in the resultset for each reklamation (one for each corresponding reklamation_status ).

Another thing is to consider is the type of the JOIN . INNER JOIN would filter out records of reklamation s that do not have a reklamation_status : the original query with the inline subquery does not behave like that, so I chose LEFT JOIN . If you can guarantee that every reklamation has at least one child in reklamation_status , you can safely switch back to INNER JOIN (which might perform more efficiently).

---

PS:

I read that subqueries are slower than JOIN s.

This is not a universal truth. It depends on many factors and cannot be told without seeing your exact use case.

Answer 3

What you read is incorrect. Subqueries can be slower, faster, or the same as joins. I would write the query as:

SELECT r.id, r.titel,
       (SELECT rs.status 
        FROM reklamation_status rs
        WHERE rs.id_reklamation = r.id 
        ORDER BY rs.id DESC 
        LIMIT 1 
       ) as status
FROM reklamation  r
WHERE r.aktiv = 1;

For performance, you want an index on reklamation_status(id_reklamation, id, status) .

By the way, this is a case where the subquery is probably the fastest method for expressing this query. If you attempt a JOIN , then you need some sort of aggregation to get the most recent status. That can be expensive.