I have three tables
Post
ID Name
1 \'Something\'
2 \'Something else\'
3 \'One more\'
Comment
ID PostId Profile
You could use a nested select like this:
SELECT Post.Id, temp.Count
FROM Post
LEFT JOIN
(SELECT Post.Id, COUNT(Comment.ID) AS Count
FROM Post
LEFT JOIN Comment ON Comment.PostId = Post.ID
LEFT JOIN Profile ON Profile.ID = Comment.ProfileID
WHERE Profile.Approved = 1
GROUP BY Post.Id)
temp ON temp.Id = Post.ID
Which would give you null where there are no posts, rather than no record:
1 1
2 null
3 1
Just to improve on that, you could use an if to get rid of the nulls
SELECT Post.Id, if(temp.Count >= 1,temp.Count,0) as newCount
FROM Post
LEFT JOIN
(SELECT Post.Id, COUNT(Comment.ID) AS Count
FROM Post
LEFT JOIN Comment ON Comment.PostId = Post.ID
LEFT JOIN Profile ON Profile.ID = Comment.ProfileID
WHERE Profile.Approved = 1
GROUP BY Post.Id) temp ON temp.Id = Post.ID
Which gives you what you originally wanted:
1 1
2 0
3 1
Note: There is most probably a more elegant solution though!!!!
From the definition of the COUNT function:
The COUNT function will only count those records in which the field in the brackets is NOT NULL.
This means that simple outer join like this would work:
SELECT Post.ID, COUNT(Comment.ID)
FROM Post LEFT JOIN Comment ON (Post.ID = Comment.PostId)
LEFT JOIN Profile ON (Profile.ID = Comment.ProfileID AND
Profile.Approved = 1)
GROUP BY Post.ID
SELECT Post.Id, COUNT(Comment.ID) AS Count
FROM Post
LEFT JOIN Comment ON Comment.PostId = Post.ID
LEFT JOIN Profile ON Profile.ID = Comment.ProfileID
WHERE Profile.Approved = 1
GROUP BY Post.Id
Probably you didn't paste it for the sake of the example, but you might evaluate to de-normalize the Profile
table together with the Comment
one, by moving the Approved
column in it.