从多列中选择 DISTINCT

Question

这是一个有点复杂的查询，但我有一个斯诺克比赛数据库，我正在尝试生成谁打过比赛并赢得了最多决定性的比赛的统计数据（如果你不担心规则，请不要担心了解游戏）。

表：

ID player1  player2   bestOf     player1Score      player2Score
1   1           2          9          5                   0
2   2           1          9          5                   4
3   1           2          9          5                   4
4   2           1          9          4                   5

我正在尝试做的事情如下：

SELECT COUNT(*) AS played, DISTINCT(player1,player2) AS playerID 
FROM matches 
WHERE (player1Score=BestOf / 2 + 0.5 AND player2Score=BestOf / 2 - 0.5) 
GROUP BY playerID

上面的查询不起作用，因为我相信 DISTINCT 不支持多列。我从上表中寻找的结果是：

playerID played  won
1           3      2
2           3      1

表格中的第一行不显示，因为它不是最后一帧。

我尝试过以下变体：

SELECT GROUP(player1,player2)
SELECT player1 + player2 AS playerID, select DISTINCT(playerID)
SELECT (player1 + player2) AS playerID GROUP BY playerID

还有一些其他的。任何提示将不胜感激！

Answer 1

我实际上并没有你的表，所以我的 sql 可能有很多语法错误。但希望我能理解这个总体思路。 我们可以计算player1获胜的决胜局列表，并将其与player2获胜的决胜局列表合并。只要我们将获胜者的列名重命名为相同的字符串（'player'），这些表的联合应该会产生一个表，标出每个人获胜的次数。

select player, count(*) as num_won from ((select player1 as player from matches where player1Score - player2Score = 1) union (select player2 as player from matches where player2Score - player1Score = 1));

这应该为您提供从每个玩家的 id 到他们获胜次数的映射。

然后考虑

select player, count(*) as num_played from ((select player1 as player from matches where abs(player1Score - player2Score) = 1) union (select player2 as player from matches where abs(player2Score - player1Score) = 1));

abs 函数的使用应该可以帮助您了解每个玩家玩的决胜局游戏的数量。现在我们将这两张表结合起来得到你的最终答案

select players_table.player, players_table.num_played, winners_table.num_won from (select player, count(*) as num_won from ((select player1 as player from matches where player1Score - player2Score = 1) union (select player2 as player from matches where player2Score - player1Score = 1)) winners_table), (select player, count(*) as num_played from ((select player1 as player from matches where abs(player1Score - player2Score) = 1) union (select player2 as player from matches where abs(player2Score - player1Score) = 1)) players_table) where winners_table.player == players_table.player;

Answer 2

create table temp_score
select playerid, count(*) as won 
FROM(
  Select   case  when player1score > player2score then  player1 else player2 end as playerid
    from score
    where
    (player2Score=Round(BestOf / 2 + 0.5,0) AND player1Score=round(BestOf / 2 - 0.5,0))
    or
    (player1Score=Round(BestOf / 2 + 0.5,0) AND player2Score=round(BestOf / 2 - 0.5,0))
) a
group by playerid

Select playerid,won, (Select SUM(won) from temp_score)  as TotalPlayed  from temp_score