Question

I’m having trouble understanding this query:

SELECT * 
  FROM `advertise` parent 
 WHERE 3  <= ( SELECT COUNT(username) 
                 FROM `advertise` 
                WHERE keyword = parent.keyword 
                  AND bid > parent.bid) 
   AND username = 'mike'

What this query does is display any rows with the username 'mike'; if that row is not in the three highest bids for a keyword.

Here is the sqlfiddle that works, but I don’t understand how it works. More specifically I don’t understand 3 <= with this specific query.

How does 3 <= determine the row is not in the three highest bids for a keyword?

Answer 1

This query works like so:

The subquery

SELECT COUNT(username) FROM `advertise` WHERE keyword = parent.keyword AND bid > parent.bid

finds the number of rows that have a higher bid for the same keyword. We then specify that the bid we are looking for cannot be in the top three bids, so we require that at least 3 bids be returned from that query. Here is what I think your end query should look like:

SELECT * FROM `advertise` parent 
WHERE 3  <= ( SELECT COUNT(username) 
    FROM `advertise` 
    WHERE keyword = parent.keyword 
    AND bid > parent.bid) 
AND username = 'mike'