如何在 SQL Server 中调整每月温度？

Posted 2023-05-08

【中文标题】如何在 SQL Server 中调整每月温度？

【英文标题】：How to pivot temperature per month in SQL Server?

【发布时间】：2020-10-09 12:35:44

【问题描述】：

我可以这样算出每年的平均温度：

select Ville, 
       AVG(Temperature) as Temperature 
from Meteo 
where DateDonnees between '2011-01-01' and '2012-01-01' 
group by  Ville

问题是我想要每个月的温度。

所以我找到了一个post，它解释了如何做到这一点：

select * from(
select Ville, Temperature as Temperature from Meteo where DateDonnees between '2011-01-01' and '2012-01-01' group by  Ville
)
pivot(
cast(AVG(Temperature)) AS DECIMAL(4, 1) for DateDonnees in (
    1 JAN, 2 FEB, 3 MAR, 4 APR, 5 MAY, 6 JUN,
    7 JUL, 8 AUG, 9 SEP, 10 OCT, 11 NOV, 12 DEC)
))

它会抛出 "Incorrect syntax near the keyword 'pivot'." 错误。

这是我希望看到的结果：

【问题讨论】：

请以表格文本形式向我们展示您想要的结果。

【参考方案1】：

MONTH 函数可以放置在横向连接中，然后在 CASE 逻辑中用于旋转月份。像这样的

select Ville as City, 
       avg(case when mo.mon=1 then Temperature else 0 end) Jan,
       avg(case when mo.mon=2 then Temperature else 0 end) Feb,
       avg(case when mo.mon=3 then Temperature else 0 end) Mar,
       avg(case when mo.mon=4 then Temperature else 0 end) Apr,
       avg(case when mo.mon=5 then Temperature else 0 end) May,
       avg(case when mo.mon=6 then Temperature else 0 end) Jun,
       avg(case when mo.mon=7 then Temperature else 0 end) Jul,
       avg(case when mo.mon=8 then Temperature else 0 end) Aug,
       avg(case when mo.mon=9 then Temperature else 0 end) Sep,
       avg(case when mo.mon=10 then Temperature else 0 end) Oct,
       avg(case when mo.mon=11 then Temperature else 0 end) Nov,
       avg(case when mo.mon=12 then Temperature else 0 end) [Dec]
from Meteo m
     cross apply
     (select month(m.DateDonnees) mon) mo
where DateDonnees between '2011-01-01' and '2012-01-01' 
group by Ville;

【参考方案2】：

我会推荐条件聚合：

select city,
    avg(case when datedonnees >= '20200101' and datedonnees < '20200201' then temperature end) jan,
    avg(case when datedonnees >= '20200201' and datedonnees < '20200301' then temperature end) feb,
    avg(case when datedonnees >= '20201201' and datedonnees < '20200401' then temperature end) mar,
    ...
    avg(case when datedonnees >= '20201201' and datedonnees < '20210101' then temperature end) dec
from meteo
where datedonnees >= '20200101' and datedonnees < '20210101'
group by city

我们可以使用month() 稍微缩短语法，但它可能比半开区间效率低：

select city,
    avg(case when month(datedonnees) =  1 then temperature end) jan,
    avg(case when month(datedonnees) =  2 then temperature end) feb,
    avg(case when month(datedonnees) =  3 then temperature end) mar,
    ...
    avg(case when month(datedonnees) = 12 then temperature end) dec
from meteo
where datedonnees >= '20200101' and datedonnees < '20210101'
group by city