SQL计算同一张表中相同ID的多列

Question

我有一个包含 GameID, GoalID, PlayerID, Assist1ID, Assist2ID 列的表（所有整数）。 PlayerID 翻译为实际进球的人的ID，但Assist1ID 和Assist2ID 也可以获得球员ID。

我正在尝试获取具有不同 PlayerID 的数据集（来自 PlayerID 列或任一助攻列）、进球数（PlayerID 列）和助攻数（其中实际上是 Assist1 和 Assist2 发生 playerID 时计数列的总和）。 PlayerID 永远不会出现在这些列中的一个以上。

我一直在尝试几种方法，主要是使用 UNION ALL，以及一些 SUM/CASE，但我就是不明白。我应该为此使用临时表，还是有办法检查行，如果PlayerID。

示例：（请注意，GoalID 和 GameID 在这种情况下并不重要）

GameID | GoalID | PlayerID   | Assist1ID | Assist2ID
  1    |    1   |     1876   |   2098    |     1097 
  1    |    2   |     2098   |   1829    |     1876 
  1    |    3   |     2098   |   1876    |     ---- 

我的查询应该返回：

PlayerID | Goals   | Assists
 1876    |    1    |    2
 2098    |    2    |    1
 1829    |    0    |    1
 1097    |    0    |    1

等

这真的可能吗，还是我必须在代码部分做一些工作？

Answer 1

为确保您获得每个参与的玩家的结果记录，无论他们是只参与、只协助还是两者都参与，您必须三次检查您的数据并将记录与 UNION ALL 粘合在一起。然后数数。

select playerid, sum(goal) as goals, sum(assist) as assists
from
(
  select playerid, 1 as goal, 0 as assist from mytable
  union all
  select assist1id, 0 as goal, 1 as assist from mytable
  union all
  select assist2id, 0 as goal, 1 as assist from mytable
)
group by playerid;  

Answer 2

可以这样做，但我感觉可能有更简单的解决方案。

SELECT PlayerID, SUM(Goals), SUM(Assists)
FROM (
    SELECT PlayerID,Count(*) AS Goals,0 AS Assists FROM Goals GROUP BY PlayerID UNION ALL
    SELECT Assist1ID,0,Count(*) FROM Goals GROUP BY Assist1ID UNION ALL 
    SELECT Assist2ID,0,Count(*) FROM Goals GROUP BY Assist2ID
) T
WHERE NOT PlayerID IS NULL
GROUP BY PlayerID

Answer 3

SELECT x.playerid
     , SUM(y.playerid = x.playerid) goals
     , SUM(x.playerid IN (y.assist1id,y.assist2id)) assists
  FROM 
     ( SELECT playerID FROM my_table 
        UNION 
       SELECT assist1id FROM my_table 
        UNION  
       SELECT assist2id FROM my_table
     ) x
  LEFT
  JOIN my_table y
    ON x.playerid IN(y.playerid,y.assist1id,y.assist2id)
 WHERE x.playerid IS NOT NULL
 GROUP
    BY playerid;