时间戳

首页 > 解决方案 > 在 SQL 中取消透视表并按条件连接

sql-server - 在 SQL 中取消透视表并按条件连接

问题描述

我有 2 张桌子：

资金

country   2002     2006    2010   2014
Chicago   1000    23000    4500   9076

奖牌

year   sport   event    country   gender   medal_rank  medal
2006    A        Z      Chicago    F          2         silver
1924    B        ZZ     Newyork    M          3         bronze
2014    C        ZZZZ   Chicago    M          1         Gold
1967    B        ZZb    Newyork    M          3         bronze
2014    C        ZZa    Chicago    F          2         silver

我想取消透视基金中的数据，例如

country    year    value
Chicago    2002    1000
Chicago    2006    23000
Chicago    2010    4500
Chicago    2014    9076

然后加入 Medals，只带来 2002、2006、2010 和 2104 年的奖牌列计数，因为我们在 Funds 中只有这些年份的值。在最后一步中，我将 value/countofmedal 划分为 value/medal。

决赛桌应如下所示：

country    year    value     countofmedeals      value/medal
Chicago    2002    1000        0                     0
Chicago    2006    23000       1                     23000
Chicago    2010    4500        0                     0
Chicago    2014    9076        2                     4536

以下是我尝试过的代码，但它不起作用。

use RDG

select
    S.Yr,
    qu.year,
    s.Cost,
    count(qu.medal)
from
(
    select 
        country,
        yr, 
        Cost
    from Funds
    unpivot
    (
        Yr for Cost in ([2002],[2006],[2010],[2014])
    ) as u
) s
join medals qu
on (s.country=qu.country and s.Yr=qu.year)
group by s.Yr,qu.year,s.Cost

标签： sql-serverjoinunpivot

解决方案

declare @Funds table
(
    country varchar(20),
    [2002] int,
    [2006] int,
    [2010] int,
    [2014] int
);
insert into @Funds(country, [2002], [2006], [2010], [2014])
values ('Chicago', 1000, 23000, 4500, 9076);


declare @Medals table
(
year int,
sport varchar(20),
event varchar(20),
country varchar(20),
gender char(1),
medal_rank tinyint,
medal varchar(20)
);
insert into @Medals(year, sport, event, country, gender, medal_rank, medal)
values
(2006, 'A', 'Z', 'Chicago', 'F', 2, 'silver'),
(1924, 'B', 'ZZ', 'Newyork', 'M', 3, 'bronze'),
(2014, 'C', 'ZZZZ', 'Chicago', 'M', 1, 'Gold'),
(1967, 'B', 'ZZb', 'Newyork', 'M', 3, 'bronze'),
(2014, 'C', 'ZZa', 'Chicago', 'F', 2, 'silver');

select
    S.country,
    S.Yr,
    qu.year /*??*/,
    s.value,
    count(qu.medal) as countofmedals,
    cast(isnull(s.value*1./nullif(count(qu.medal), 0), 0) as decimal(19,2)) as [value/medal]
from
(
    select 
        country,
        yr, 
        value
    from @Funds
    unpivot
    (
        value for yr in ([2002],[2006],[2010],[2014]) --!!!!
    ) as u
) s
left join @medals qu --!!!!
on (s.country=qu.country and s.Yr=qu.year)
group by s.country, s.Yr,qu.year/*??*/,s.value;

推荐阅读