join - SQL query: Simulating an "AND" over several rows instead of sub-querying

Question

Suppose I have a "tags" table with two columns: tagid and contentid. Each row represents a tag assigned to a piece of content. I want a query that will give me the contentid of every piece of content which is tagged with tagids 334, 338, and 342.

The "easy" way to do this would be (pseudocode):

select contentid from tags where tagid = 334 and contentid in (
    select contentid from tags where tagid = 338 and contentid in (
        select contentid from tags where tagid = 342
    )
)

However, my gut tells me that there's a better, faster, more extensible way to do this. For example, what if I needed to find the intersection of 12 tags? This could quickly get horrendous. Any ideas?

EDIT: Turns out that this is also covered in this excellent blog post.

Answer 1

SELECT contentID
FROM tags
WHERE tagID in (334, 338, 342)
GROUP BY contentID
HAVING COUNT(DISTINCT tagID) = 3


--In general
SELECT contentID
FROM tags
WHERE tagID in (...) --taglist
GROUP BY contentID
HAVING COUNT(DISTINCT tagID) = ... --tagcount