Question

我有一个“积分榜”查询，显示我所在的联赛的统计数据。我试图根据一些条件订购查询。

如果团队的积分排名为1-25，则按胜利排序
如果球队排名前25位（积分排在第26位至第42位），请按积分排序。

我能描述它的最好方法是它应该像两张桌子一样。排名（1-25）的队伍应该排在一起，排名（26-42）的队伍应该排在一起。

这是我的查询。

SELECT m.Team AS team
, SUM( r.points ) AS points
, SUM(CASE
    WHEN rank = 1 THEN 1
    ELSE 0
  END) AS Wins
, SUM(CASE
    WHEN rank < 6 THEN 1
    ELSE 0
  END) AS T5
, SUM(CASE
    WHEN rank < 11 THEN 1
    ELSE 0
  END) AS T10
, SUM(CASE
    WHEN rank < 21 THEN 1
    ELSE 0
  END) AS T20
, ROUND(Avg(r.points),2) AS ppr
, ROUND(Avg(r.rank),2) as avg_finish
, MIN(r.rank) as best
FROM members m
LEFT JOIN results r ON r.team_id = m.M_ID
GROUP BY team
ORDER BY 
    CASE
        WHEN COUNT(*) < 26
        THEN Wins 
        WHEN COUNT(*) > 25
        THEN points
    END DESC, points DESC;

由于我的别名“Wins”，我收到错误“引用'Wins'不支持（引用群组功能）。相反，我尝试过：

ORDER BY 
    CASE
        WHEN COUNT(*) < 26
        THEN SUM(CASE
            WHEN rank = 1 THEN 1
            ELSE 0
        END)  
        WHEN COUNT(*) > 25
        THEN points
    END DESC, points DESC;

这成功订购了我的桌子，但没有考虑COUNT（*）＆gt; 25 CASE条款，只是按顺序，然后点。关于如何重构ORDER BY子句的任何想法？

Answer 1

您应该在外部查询中执行ORDER BY，因为Wins子句中无法访问列别名ORDER。

SELECT * FROM (
SELECT m.Team AS team
, SUM( r.points ) AS points
, SUM(CASE
    WHEN rank = 1 THEN 1
    ELSE 0
  END) AS Wins
, SUM(CASE
    WHEN rank < 6 THEN 1
    ELSE 0
  END) AS T5
, SUM(CASE
    WHEN rank < 11 THEN 1
    ELSE 0
  END) AS T10
, SUM(CASE
    WHEN rank < 21 THEN 1
    ELSE 0
  END) AS T20
, ROUND(Avg(r.points),2) AS ppr
, ROUND(Avg(r.rank),2) as avg_finish
, MIN(r.rank) as best
FROM members m
LEFT JOIN results r ON r.team_id = m.M_ID
GROUP BY team ) XXX
ORDER BY 
    CASE
        WHEN COUNT(*) < 26
        THEN Wins 
        WHEN COUNT(*) > 25
        THEN points
    END DESC, points DESC;

Answer 2

对Wins的引用应该有效。 MySQL接受order by中的列别名，因此不需要子查询。

但是，您可能希望逻辑为：

ORDER BY (CASE WHEN COUNT(*) < 26 THEN 1 ELSE 2 END),  -- put the top ranking first,
         (CASE WHEN COUNT(*) < 26 THEN wins END) DESC,
         points DESC

在MySQL中，您还可以简化select逻辑：

SELECT m.Team AS team, SUM( r.points ) AS points,
       SUM(rank = 1) then Wins,
       SUM(rank < 6) as T5,
       SUM(rank < 11) as T10,
       SUM(rank < 21) as T20
       . . .

编辑：

我认为您想要的ORDER BY是：

ORDER BY (points <= 26) DESC,  -- put the top ranking first,
         wins desc,
         points DESC

Answer 3

不幸的是，MySQL不支持ROW_NUMBER()或RANK()等分析函数。如果确实如此，我们可以使用ROW_NUMBER()函数根据特定订单识别“前25行”。但是有两种方法可以模拟ROW_NUMBER()函数。

一种方法是使用带有用户定义变量的技巧来模拟它。我们使用这种方法是我们需要标记行，并返回值。

在这种情况下，我们只关心一支球队是否“进入前25名”。所以我们可以运行一个单独的查询来返回“前25名”团队。

例如，此查询获取两个集......“所有团队”（a）的集合结果以及“前25名”团队的集合（{{ 1}}），并匹配使用外连接操作的那些：

内联视图“a”是获取所有团队的查询。

内联视图“t”本质上是相同的查询，但有一个ORDER BY和LIMIT子句，因此它最多返回25行。 ORDER BY在LIMIT之前应用，因此它将返回具有最高分数的25支球队。（其他表达式包含在ORDER BY中，以使结果更具确定性......如果列表中第25和第26的两个团队具有相同的点数......

外部查询识别来自SELECT a.* FROM ( SELECT m.team AS team , SUM( r.points ) AS points , SUM(IF( r.rank=1 ,1,0)) AS wins , SUM(IF( r.rank<6 ,1,0)) AS t5 , SUM(IF( r.rank<11 ,1,0)) AS t10 , SUM(IF( r.rank<21 ,1,0)) AS t20 , ROUND(AVG( r.points ),2) AS ppr , ROUND(AVG( r.rank ),2) AS avg_finish , MIN( r.rank ) AS best FROM members m LEFT JOIN results r ON r.team_id = m.m_id GROUP BY m.team ) a LEFT JOIN ( SELECT tm.team AS team , SUM(IF( tr.rank=1 ,1,0)) AS wins FROM members tm LEFT JOIN results tr ON tr.team_id = tm.m_id GROUP BY tm.team ORDER BY SUM( tr.points ) DESC , SUM(IF( tr.rank=1 ,1,0)) DESC , tr.team_id DESC LIMIT 25 ) t ON t.team = ta.team ORDER BY t.wins DESC , a.points DESC , a.team DESC的哪个团队位于“前25名”中，并应用适当的排序。（请注意，没有来自a的匹配行的团队将为t中的所有列提供NULL值...即t将为NULL，因此ORDER BY可以正常工作。

我们不一定需要从内嵌视图t.wins返回wins。我们可以通过测试NULL值来检查是否从t返回了匹配的行。例如

Answer 4

您的问题是，在按points对结果进行排序之前，您不知道排名。如果你知道等级，你可以用

排序

ORDER BY (rank > 25) ASC
, CASE WHEN rank <= 25 THEN Wins ELSE 0 END
, points

确定MySQL排名的一种方法是将排序结果存储在具有AUTO_INCREMENT排名列的临时表中。

DROP TEMPORARY TABLE IF EXISTS tmp_teams;
CREATE TEMPORARY TABLE tmp_teams (
    rank INT AUTO_INCREMENT PRIMARY KEY,
    team varchar(255),
    points INT,
    Wins INT,
    T5 INT,
    T10 INT,
    T20 INT,
    ppr INT,
    avg_finish INT,
    best INT
) AS
    SELECT null AS rank
    , m.Team AS team
    , SUM( r.points ) AS points
    , SUM(CASE
        WHEN rank = 1 THEN 1
        ELSE 0
      END) AS Wins
    , SUM(CASE
        WHEN rank < 6 THEN 1
        ELSE 0
      END) AS T5
    , SUM(CASE
        WHEN rank < 11 THEN 1
        ELSE 0
      END) AS T10
    , SUM(CASE
        WHEN rank < 21 THEN 1
        ELSE 0
      END) AS T20
    , ROUND(Avg(r.points),2) AS ppr
    , ROUND(Avg(r.rank),2) as avg_finish
    , MIN(r.rank) as best
    FROM members m
    LEFT JOIN results r ON r.team_id = m.M_ID
    GROUP BY team
    ORDER BY points DESC, Wins DESC, r.team_id ASC -- define the order for rank
;

SELECT *
FROM tmp_teams
ORDER BY (rank > 25) ASC
, CASE WHEN rank <= 25 THEN Wins ELSE 0 END
, points;

另一种方法是使用会话变量（@rank），它将在有序子查询中递增：

SELECT *
FROM (
    SELECT m.Team AS team
    , SUM( r.points ) AS points
    , SUM(CASE
        WHEN rank = 1 THEN 1
        ELSE 0
      END) AS Wins
    , SUM(CASE
        WHEN rank < 6 THEN 1
        ELSE 0
      END) AS T5
    , SUM(CASE
        WHEN rank < 11 THEN 1
        ELSE 0
      END) AS T10
    , SUM(CASE
        WHEN rank < 21 THEN 1
        ELSE 0
      END) AS T20
    , ROUND(Avg(r.points),2) AS ppr
    , ROUND(Avg(r.rank),2) as avg_finish
    , MIN(r.rank) as best
    , @rank := @rank + 1 AS rank
    FROM members m
    LEFT JOIN results r ON r.team_id = m.M_ID
    CROSS JOIN (select @rank := 0) AS init_rank
    GROUP BY team
    ORDER BY points DESC, Wins DESC, r.team_id ASC -- define the order for rank
    LIMIT 1000000000 -- workaround for some versions/settings to force sorting a subquery result
) sub
ORDER BY (rank > 25) ASC
, CASE WHEN rank <= 25 THEN Wins ELSE 0 END
, points

Answer 5

它终于奏效了。保罗，你的排名是正确的。以下是Mahmoud提供我使用的排名的链接： Rank Values

SET @rownum = 0; 

SELECT *, (@rownum := @rownum + 1) AS chase FROM (
SELECT m.Team AS team
, SUM( r.points ) AS points
, SUM(r.rank = 1) AS Wins
, SUM(r.rank < 6) AS T5
, SUM(r.rank < 11) AS T10
, SUM(r.rank < 21) AS T20
, ROUND(Avg(r.points),2) AS ppr
, ROUND(Avg(r.rank),2) as avg_finish
, MIN(r.rank) as best
FROM members m
LEFT JOIN results r ON r.team_id = m.M_ID
GROUP BY team
ORDER BY points DESC ) AS ptotal
ORDER BY CASE
    WHEN chase <= 24
    THEN Wins
END DESC, points DESC;

谢谢大家的帮助。我刚刚开始使用这个网站，无法相信你们都有多大帮助。当我在路上遇到问题时，我知道在哪里看看！

使用CASE MySQL

5 个答案: