Question

我正在尝试选择一个表，显示在SQL Server中具有特定ID的某些列上执行的操作的结果。

EventLog表包含代表不同游戏事件的各种EventId（17 =死亡，21 =杀死）。我正在尝试获取死亡事件的FloatResult列的SUM和AVG，以及kill事件的记录总数，然后将它们显示在表与类类型上。

select 
    e.Class, a.TotalLifetime, a.AverageLifetime, b.TotalKills
from 
    EventLog e 
join 
    (select 
         e1.Class, 
         SUM(e1.FloatResult) as TotalLifetime, 
         AVG(e1.FloatResult) as AverageLifetime
     from 
         EventLog e1
     join 
         Maps m on e1.MapId = m.Id
     where 
         e1.EventId = 17 and
         m.MapName like '%my_map' and 
         e1.EventTime >= '2013-07-04' and
         e1.EventTime <= DATEADD(d, 1, '2013-07-05')
     group by 
         e1.Class) as a on e.Class = a.Class
join (
    select e2.Class, count(*) as TotalKills
    from EventLog e2 
    join Maps m on e2.MapId = m.Id
    where e2.EventId = 21 and
    m.MapName like '%my_map' and 
        e2.EventTime >= '2013-07-04' and
        e2.EventTime <= DATEADD(d, 1, '2013-07-05')
    group by e2.Class
) as b on e.Class = b.Class
group by e.Class

理想的结果是：

Class  TotalLifeTime AverageLifetime TotalKills
0      563.45        30.5            100
1      766.6         12.56           20
etc...

这导致SQL Server说：

错误第1行：列'a.TotalLifetime'在选择列表中无效，因为它不包含在聚合函数或GROUP BY子句中。

但是我不想按照这个值进行分组，只能按类进行分组。

有没有更好的方法来制定此查询？

Answer 1

您要按class预先汇总所有数据字段。因此，您不需要在最外层再次聚合。您只需要从EventLog获取不同的类：

select 
    e.Class, a.TotalLifetime, a.AverageLifetime, b.TotalKills
from 
    (select distinct class from EventLog e
    ) e
left outer join 
    (select 
         e1.Class, 
         SUM(e1.FloatResult) as TotalLifetime, 
         AVG(e1.FloatResult) as AverageLifetime
     from 
         EventLog e1
     join 
         Maps m on e1.MapId = m.Id
     where 
         e1.EventId = 17 and
         m.MapName like '%my_map' and 
         e1.EventTime >= '2013-07-04' and
         e1.EventTime <= DATEADD(d, 1, '2013-07-05')
     group by 
         e1.Class) as a on e.Class = a.Class
left outer join (
    select e2.Class, count(*) as TotalKills
    from EventLog e2 
    join Maps m on e2.MapId = m.Id
    where e2.EventId = 21 and
    m.MapName like '%my_map' and 
        e2.EventTime >= '2013-07-04' and
        e2.EventTime <= DATEADD(d, 1, '2013-07-05')
    group by e2.Class
) as b on e.Class = b.Class;

我还将join更改为left outer join以保留所有课程。

数学运算来自同一个表，具体取决于具体的id

1 个答案: