Question

I have two queries to get related tags from a mysql database, one works, one does not, my question is: “why?”

Problem:
When executing the first query, the mysql server gets 100% cpu usage, and has to be restarted to function again.

Query 1 (does not work):

SELECT tags.*, COUNT(ct.company_id) AS count
FROM company2tag ct, tags
WHERE ct.company_id IN (
    SELECT ct.company_id FROM company2tag ct
    WHERE ct.tag_id = 18
    GROUP BY ct.company_id
    HAVING COUNT(ct.company_id) = 1
)
AND tags.id != 18
AND tags.id = ct.tag_id
GROUP BY ct.tag_id
ORDER BY count DESC
LIMIT 5;

Query 2 (works):

SELECT tags.*, COUNT(ct.company_id) AS count
FROM company2tag ct, tags
WHERE ct.company_id IN (5864, 5870, 6140, 6221, 6268)
    AND tags.id != 18
    AND tags.id = ct.tag_id
GROUP BY ct.tag_id
ORDER BY count DESC
LIMIT 5;

To my understanding the two queries above do completely the same, the only difference is that the first query retrieves its “company_id’s” via a subselect.

How can this happen?

Answer 1

First of all, you may be experiencing problems from the first query because you have two tables aliased to ct… one in the outer query, one in the sub-query.

Secondly, you can rewrite the IN as a JOIN:

SELECT tags.*, COUNT(ct.company_id) AS count
FROM company2tag ct
INNER JOIN tags ON tags.id = ct.tag_id
INNER JOIN (
    SELECT company_id FROM company2tag
    WHERE tag_id = 18
    GROUP BY company_id
    HAVING COUNT(company_id) = 1
) ctf ON ct.company_id = ctf.company_id
WHERE tags.id != 18
GROUP BY ct.tag_id
ORDER BY count DESC
LIMIT 5;

Note that I haven’t actually tested this.