MySql LeftJoin从三个表

Question

I am trying to get the data from the game table and whether all gamers have signed up for an available match(Yes/No) An available match is one where the date is current or in the future

Gamers

(PK)Gamer_Id
Gamer_firstName,
Gamer_lastName,
Gamer experience(Y/N) 

Game_matches

(PK)FK GamerId,
(PK)FK MatchId,
 Gamer_score

Match

(PK)Match_Id,
gameDate,
ExperiencedGamers_needed,
InExperiencedGamers_needed

This is what I've tried, I know it's incomplete. Do I need to use a CASE as well?

SELECT G.*, COUNT(M.`matchId`)
FROM game_matches GM
LEFT JOIN MATCHES M
ON M.`matchId` = GM.`matchId`
LEFT JOIN Gamers G ON G.`userId` = GM.`userId`
WHERE G.`gameDate` >= DATE(NOW())
GROUP BY G.`userId`

Are there any other ways to acheive this other than using the cross join solution from JW?

Answer 1

SELECT  a.GamerID,
        a.Name,
        COUNT(*) = COUNT(c.GamerID)
FROM    Gamers a
        CROSS JOIN `Match` b
        LEFT JOIN GameMatches c
            ON  a.GamerID = c.GamerID AND
                b.MatchID = c.MatchID
WHERE   b.gameDate >= CURDATE()
GROUP   BY  a.GamerID,
            a.Name

• SQLFiddle Demo

What the query does is it first gets the cartesian product of all gamers and all the matches starting today. After that, the results will be joined on table GameMatches on two columns: GamerID and MatchID .

If the result of the cartesian product of the tables has no matches on table GameMatches , the values of the columns are NULL . Thus, it test if COUNT(*) which is the total number of records is equal to COUNT(c.GamerID) which is only the total number of gamers that have signed up.

Answer 2

You are only selecting the games that have matches in the future. If you want the count of future matches, you need a conditional aggregation:

SELECT G.*, sum(case when g.gameDate >= date(now()) then 1 else 0 end) as FutureMatches
FROM game_matches GM
LEFT JOIN MATCHES M
ON M.`matchId` = GM.`matchId`
LEFT JOIN Gamers G ON G.`userId` = GM.`userId`
GROUP BY G.`userId`

In practice, you may be able to use inner join rather than left join because the ids should match between tables (unless you are interested in users that have no matches at all).

Answer 3

WITH MATCHES AS
(
SELECT COUNT(*) AS TOTAL_MATCHES FROM MATCH WHERE gameDate >= CURR_DATE
)

GM AS
(SELECT G.GAMER_ID, COUNT(*) AS GAMER_MATCHES
GAMERS G, GAME_MATCHES GM
WHERE G.GAMER_ID = GM.GAMER_ID
GROUP BY G.GAMER_ID
)

SELECT GM.GAMER_ID
FROM GM, MATCH
WHERE GAMER_MATCHES = TOTAL_MATCHES