MySQL联接查询结果按标签过滤

Question

I am struggling with one MySQL query and I can figure it out. I have three tables containing info about url addresses.

One, 'domains', stores unique domain names, which is 'parent' table to the second table, 'urls', storing unique urls name with some statistics. To each url record can be added multiple tags, which are stored in 'tags' table and there is of course connecting table between urls and tags called 'url_has_tag'.

I have query that selects domain name, url stats and concatenated tags for all url records:

SELECT u.url_id, u.url, d.domain, u.hits, u.version, GROUP_CONCAT( t.label) as tags
     FROM urls AS u
     INNER JOIN domains AS d 
        ON (u.parent_id = d.domain_id) 
    LEFT JOIN url_has_tag AS ut ON u.url_id = ut.url_id 
    LEFT JOIN tags as t ON ut.tag_id = t.tag_id
GROUP BY u.url_id
ORDER BY u.hits DESC

Now I need to add filtering results by tags, which means I want to select select the same information but only for urls which are tagged with some tag. They can be tagged by other too, and I want to have the other tags included in the result.

When I added WHERE clause before grouping WHERE ut.tag_id = 7 it selects right rows, but the other tags are not included (as expected).

Could anyone help me with this? Thank you very much.

Answer 1

By adding where clause on right side table of left outer join makes it inner join. So add the ut.tag_id = 7 in the ON clause of join as below.

SELECT u.url_id, u.url, d.domain, u.hits, u.version, GROUP_CONCAT( t.label) as tags
 FROM urls AS u
 INNER JOIN domains AS d 
    ON (u.parent_id = d.domain_id) 
LEFT JOIN url_has_tag AS ut ON u.url_id = ut.url_id and ut.tag_id = 7
LEFT JOIN tags as t ON ut.tag_id = t.tag_id
GROUP BY u.url_id
ORDER BY u.hits DESC

Answer 2

Solved thanks to a friend!

SELECT u.url_id, u.url, d.domain, u.hits, u.version, GROUP_CONCAT(DISTINCT(t2.label)) as tags
     FROM urls AS u
     INNER JOIN domains AS d
        ON (u.parent_id = d.domain_id)
     LEFT JOIN url_has_tag AS ut ON u.url_id = ut.url_id
     LEFT JOIN url_has_tag AS ut2 ON u.url_id = ut2.url_id
     LEFT JOIN tags as t ON ut.tag_id = t.tag_id
     LEFT JOIN tags as t2 ON ut2.tag_id = t2.tag_id
WHERE t.tag_id IN (7,8)
GROUP BY u.url_id
ORDER BY u.hits DESC

Tags names have to be DISTINCT, because when url matches several tags from WHERE .. IN clause, its tag names would be concatenated multiple times.