MYSQL 将一个表中的一列连接到另一个表中的两个 count()

Question

我有一个表，用于存储在线游戏的排名。该表(玩家)有 4 列(Player_ID、Player_Name、Current_ELO 和 Player_Status)。我还想显示赢/输的比赛，但该信息位于重播表中，每场比赛都有一行。 我可以使用 count(*) 函数轻松计算赢局或输局的数量:

select Winner, count(*) from Replays group by Winner;
select Loser, count(*) from Replays group by Loser;

但是，我找不到连接两个表的方法，所以我的最终结果如下:

+-------------+-------------+-----------+------------+
| Player_Name | Current_ELO | Won games | Lost games |
+-------------+-------------+-----------+------------+
| John        | 1035        | 5         | 3          |
+-------------+-------------+-----------+------------+

我最好的猜测是这样，但是返回了太多结果，因为我无法针对每个计数将“count()”按获胜者/失败者单独分组:

select Player_Name,  Current_ELO, Player_Status, count(Winner_Replays.Winner), count(Loser_Replays.Loser) from Players
inner join Replays as Winner_Replays on Winner_Replays.Winner = Players.Player_Name
inner join Replays as Loser_Replays on Loser_Replays.Loser = Players.Player_Name
group by Winner_Replays.Winner order by Current_ELO desc;

理想情况下，由于某些玩家可能取得 0 胜或 0 负，我希望它为这些用户返回 0(如果可能的话)。谢谢!

Answer 1

您可以使用以下内容，在玩家列表上使用 INNER JOIN 及其重播结果。您可以将 SUM 与 CASE 结合使用来获取每个玩家的胜利/失败次数。

SELECT Players.Player_Name, ANY_VALUE(Current_ELO), ANY_VALUE(Player_Status),
  SUM(CASE WHEN result = 'L' THEN 1 ELSE 0 END) AS Lost_Games, 
  SUM(CASE WHEN result = 'W' THEN 1 ELSE 0 END) AS Won_Games
FROM Players INNER JOIN (
  SELECT 'W' AS result, Winner AS Player_Name FROM Replays
  UNION ALL
  SELECT 'L', Loser FROM Replays
) win_lose ON Players.Player_Name = win_lose.Player_Name
GROUP BY Players.Player_Name
ORDER BY ANY_VALUE(Current_ELO) DESC;

您还可以使用 LEFT JOIN 解决方案，如下所示:

SELECT Players.Player_Name, ANY_VALUE(Current_ELO) AS Current_ELO, ANY_VALUE(Player_Status) AS Player_Status,
  COUNT(DISTINCT rLoser.ReplayID) AS Lost_Games, 
  COUNT(DISTINCT rWinner.ReplayID) AS Won_Games
FROM Players
  LEFT JOIN Replays AS rWinner ON rWinner.Winner = Players.Player_Name
  LEFT JOIN Replays AS rLoser ON rLoser.Loser = Players.Player_Name
GROUP BY Players.Player_Name
ORDER BY ANY_VALUE(Current_ELO) DESC;

demo on dbfiddle.uk