如何选择多个 GROUP BY 行的总和？

Question

我有一个球员与比赛的关系表以及他们在每场比赛中的得分。我正在尝试做一个选择，在其中我得到一个不同的球员名单，他们的总得分以及他们参加过的所有球队的总得分。由于一切都是针对单个玩家的，我不知道如何仅针对单个列摆脱 GROUP BY 范围。对于下面的例子，我只会说每支球队只有两名球员。在实际的数据库中，如果重要的话，每个团队都有五个。谢谢大家。

表“匹配”：

match_id | winning_team |

56427859 |            0 |
56427860 |            1 |
56427861 |            1 |
56427862 |            0 |
56427863 |            1 |

etc...

表格“match_players”：

match_id | team | player_id | points |

56427859 |    0 |        10 |      3 |
56427859 |    0 |        33 |      1 |
56427859 |    1 |        26 |      0 |
56427859 |    1 |        39 |      2 |
56427860 |    0 |        23 |      1 |
56427860 |    0 |        33 |      3 |
56427860 |    1 |        18 |      1 |
56427860 |    1 |        10 |      4 |

etc...

期望的结果：

player_id | match_count | total_points | team_total_points | <- This should be
                                                                the total of all
       10 |           2 |            7 |                 9 |    points scored by
       18 |           1 |            1 |                 5 |    the player and
       23 |           1 |            1 |                 4 |    his teammates
       26 |           1 |            0 |                 2 |    in all matches.
       33 |           2 |            4 |                 8 |
       39 |           1 |            2 |                 2 |

查询：

SELECT
    p.player_id,
    COUNT(*) AS match_count,
    SUM(CASE WHEN mp.team = m.winning_team THEN 1 ELSE 0 END) AS win_count,
    SUM(points) AS total_points,
    [________________________________________] AS team_total_points
FROM matches m
INNER JOIN match_players mp ON m.match_id = mp.match_id
INNER JOIN players p ON mp.player_id = p.player_id
GROUP BY player_id
ORDER BY player_id

编辑：

“球队”列简单地定义了红色或蓝色、主场或客场等。球员可以在不同的比赛中加入不同的球队。并且玩家可以在比赛之间交换球队，比如课间躲避球。

Answer 1

以下查询将计算与每个玩家在同一团队中的所有玩家的总分。

SELECT p1.player_id, SUM(p2.total_points) AS team_total_points
FROM match_players AS p1
JOIN (SELECT match_id, team, SUM(points) as total_points
      FROM match_players
      GROUP BY match_id, team) AS p2
ON p1.match_id = p2.match_id AND p1.team = p2.team
GROUP BY p1.player_id

然后您可以将其与原始查询结合起来以添加团队总数。

SELECT
    p.player_id,
    COUNT(*) AS match_count,
    SUM(CASE WHEN mp.team = m.winning_team THEN 1 ELSE 0 END) AS win_count,
    SUM(points) AS total_points,
    mp2.team_total_points
FROM matches m
INNER JOIN match_players mp ON m.match_id = mp.match_id
INNER JOIN players p ON mp.player_id = p.player_id
INNER JOIN 
    (SELECT p1.player_id, SUM(p2.total_points) AS team_total_points
     FROM match_players AS p1
     JOIN (SELECT match_id, team, SUM(points) as total_points
           FROM match_players
           GROUP BY match_id, team) AS p2
     ON p1.match_id = p2.match_id AND p1.team = p2.team
     GROUP BY p1.player_id) AS mp2 ON mp2.player_id = p.player_id
GROUP BY player_id
ORDER BY player_id

DEMO

Answer 2

试试这个：

SELECT player_id, COUNT(*) AS match_count, SUM(points) AS total_points, team_tot_points
FROM match_players JOIN (SELECT SUM(points) AS team_tot_points, team
                         FROM match_players
                         GROUP BY team) AS team_points
ON match_players.team = team_points.team
GROUP BY player_id