SQL排序基于各种列的总和

Question

我有一个MS-SQL游戏结果表，其中包含主队，主队得分，客队和客队得分。 我试图查询首先排序的队伍，胜利和损失，总胜利（降序）然后是总损失（升序）。 该表不包含赢或输列。 必须根据分数计算胜利和损失。 如果我需要创建视图或临时表，如何使用团队，胜利和损失创建该视图？

Answer 1

假设您可以在查询中提取该信息，则可以在ORDER BY子句中使用聚合：

ORDER BY
    SUM(win) DESC
   ,SUM(loss) ASC

编辑 ：更新评论

只有存储的分数，您将要么使用临时表或CTE（取决于您使用的DBMS，如果支持它们）首先创建一个包含团队及其中的胜负的表，以及然后你可以从那里SELECT和相应的ORDER 。

Answer 2

SELECT *
FROM MyTable
WHERE ...
ORDER BY SUM(win) DESC, SUM(loss) ASC

有关更详细的答案，请提供架构和示例数据。

Answer 3

家庭作业？

无论如何，您可以通过跟随订单的列来指定SQL语句中的排序顺序。 例如，

select x, y, z from table order by x asc, y desc, z asc

Answer 4

select 
    teamName, 
    'wins' = 
        case 
            when wins is null then 0 
            else wins 
        end, 
    'losses' = 
        case 
            when losses is null then 0 
            else losses 
        end 
from 
    teams 
left join 
( 
    select 
        a.wins as wins, 
        b.losses as losses, 
        'team' = 
            case 
                when a.team is null then b.team 
                else a.team 
            end 
    from 
    ( 
        select 
            homeTeam as team, 
            sum(wins) as wins 
        from 
        ( 
            select 
                homeTeam, 
                count(*) as wins 
            from 
                results 
            where 
                homeScore > awayScore 
            group by 
                homeTeam 
            union all 
            select 
                awayTeam, 
                count(*) as wins 
            from 
                results 
            where 
                homeScore < awayScore 
            group by 
                awayTeam         
        ) a 
        group by homeTeam 
    ) a 
    full outer join 
    ( 
        select 
            homeTeam as team, 
            sum(losses) as losses 
        from 
        ( 
            select 
                homeTeam, 
                count(*) as losses 
            from 
                results 
            where 
                homeScore < awayScore 
            group by 
                homeTeam 
            union all 
            select 
                awayTeam, 
                count(*) as losses 
            from 
                results 
            where 
                homeScore > awayScore 
            group by 
                awayTeam 
        ) a 
        group by 
            homeTeam 
    ) b 
    on a.team = b.team 
) c 
on teams.id = c.team