Sql server 正在尝试在SQL Server中进行求和（最大）计算

感谢你们提供的所有帮助，尽管这让我大开眼界，但不幸的是，它并没有产生我所期望的结果。为了更好地获得我所寻求的帮助，我将尝试解释我所希望实现的目标

我认为关注的主要栏目是“IN”、“AA_Now”、“STF_Now”、“dbo.Sheet1$.LOB_name”、“dbo.Sheet1$.LifeCycleName”和“dbo.Sheet1$.Area of Business”。每个“IN”都有一个“AA_Now”和“STF_Now”。一组“IN”在“dbo.Sheet1$.LOB_name”下汇总。在“dbo.Sheet1$.LOB_name”下，我只需要汇总的“IN”组的最大值。现在，“dbo.Sheet1$.LOB_name”被汇总到“dbo.Sheet1$.LifeCycleName”下，我想要的是“dbo.Sheet1$.LOB_name”下汇总的最大值的总和，以显示在“dbo.Sheet1$.LifeCycleName”的汇总中。最后，“dbo.Sheet1$.LifeCycleName”汇总到“dbo.Sheet1$.AreaOfBusiness”。和前面一样，我要寻找的是要显示的“dbo.Sheet1$.LifeCycleName”的总和。这些仅适用于“AA_Now”和“STF_Now”列 我尝试从数据透视表中进行此操作，但没有效果，我认为最好在原始数据中进行排序

我试图在SQLServer中进行求和（MAX）计算，并在执行命令时得到以下错误

味精130，15级，状态1，第6行 无法对包含聚合或子查询的表达式执行聚合函数

我确信这个错误是由这两个原因造成的

，和（最大值（转换（浮动，替换（[AA_Now]，'N/A'，'0'））为[AA2_Now]

及

，和（最大值（转换（浮动，替换（[STF_Now]，'N/A'，'0'））为[STF2_Now]

但是我不知道如何重写它而不引起错误

下面是完整的代码

SELECT  dbo.CCA_Merged.id, dbo.CCA_Merged.timeStamp, dbo.CCA_Merged.name, dbo.CCA_Merged.lN
        ,dbo.CCA_Merged.type, dbo.CCA_Merged.id2, dbo.CCA_Merged.aG 
        ,dbo.CCA_Merged.regionId, dbo.CCA_Merged.sgcc       
        ,convert(float,replace([SLC_Today],'N/A','0')) As [SLC_Today]
        ,convert(float,replace([AA_Now],'N/A','0')) As [AA_Now]
        ,SUM(MAX(convert(float,replace([AA_Now],'N/A','0')))) As [AA2_Now]
        ,convert(float,replace([SLCO_Today],'N/A','0')) As [SLCO_Today]
        ,convert(float,replace([CABN_Today],'N/A','0')) As [CABN_Today]
        ,convert(float,replace([COF_Today],'N/A','0')) As [COF_Today]
        ,convert(float,replace([HT_Today],'N/A','0')) As [HT_Today]
        ,convert(float,replace(replace([CH_Today],'N/A','0'),'-','0')) As [CH_Today]
        ,convert(float,replace([SLC_Now],'N/A','0')) As [SLC_Now]
        ,convert(float,replace([SLCO_Now],'N/A','0')) As [SLCO_Now]
        ,convert(float,replace([SLC_Thirty],'N/A','0')) As [SLC_Thirty]
        ,convert(float,replace(replace([SLCO_Thirty],'N/A','0'),'-','0')) As [SLCO_Thirty]
        ,convert(float,replace([ACWT_Today],'N/A','0')) As [ACWT_Today]
        ,convert(float,replace([CQ_Now],'N/A','0')) As [CQ_Now]
        ,convert(float,replace([LCQ_Now],'N/A','0')) As [LCQ_Now]
        ,convert(float,replace([SLCH_Now],'N/A','0')) As [SLCH_Now]
        ,convert(float,replace([STF_Now],'N/A','0')) As [STF_Now]
        ,SUM(MAX(convert(float,replace([STF_Now],'N/A','0')))) As [STF2_Now]
        ,dbo.Sheet1$.AreaOfBusiness, dbo.Sheet1$.LifeCycleName, dbo.Sheet1$.LOB_name
FROM            dbo.Sheet1$ RIGHT OUTER JOIN
                     dbo.CCA_Merged ON dbo.Sheet1$.Skill_Name = dbo.CCA_Merged.lN   
Group by ROLLUP (stf_now) ,dbo.CCA_Merged.id, dbo.CCA_Merged.timeStamp, dbo.CCA_Merged.name, dbo.CCA_Merged.lN
            ,dbo.CCA_Merged.type, dbo.CCA_Merged.id2, dbo.CCA_Merged.aG ,dbo.CCA_Merged.regionId
            ,dbo.CCA_Merged.sgcc,AA_Now,SLC_Today,SLCO_Today,CABN_Today,COF_Today,HT_Today,CH_Today
            ,SLC_Now,SLCO_Now,SLC_Thirty,SLCO_Thirty,ACWT_Today,CQ_Now,LCQ_Now,SLCH_Now
            ,dbo.Sheet1$.AreaOfBusiness, dbo.Sheet1$.LifeCycleName, dbo.Sheet1$.LOB_name

我是SQL Server的新手，非常感谢您的帮助

提前谢谢

更新的精简脚本

SELECT  dbo.CCA_Merged.lN       
    ,convert(float,replace([STF_Now],'N/A','0')) As [STF_Now]
    ,dbo.Sheet1$.LOB_name, dbo.Sheet1$.LifeCycleName, dbo.Sheet1$.AreaOfBusiness 
FROM            dbo.Sheet1$ RIGHT OUTER JOIN
                 dbo.CCA_Merged ON dbo.Sheet1$.Skill_Name = dbo.CCA_Merged.lN                   
Group by stf_now ,AA_Now,dbo.CCA_Merged.lN,dbo.Sheet1$.AreaOfBusiness, dbo.Sheet1$.LifeCycleName, dbo.Sheet1$.LOB_name
Order by AreaOfBusiness DESC

+----+---------+----------+---------------+----------------+ |在STF Now LOB name LifeCycleName业务领域| +----+---------+----------+---------------+----------------+ |A | 46 | BSW | BS |业务| |B | 46 | BSW | BS |业务| |C | 0 | BOSS | BS |业务| |D | 112 | MSD | BS |业务| |电子商务| |F | 42 | BHV | BR |业务| |G | 23 | BCR | BR |业务| |H | 23 | BHV | BR |业务| |I | 55 | BSW2 | BS |业务| |J | 1 | BSW2 | BS |业务| |K | 46 | BSW | BS |业务| |L | 112 | MSD | BS |业务| |M | 112 | MSD | BS |业务| |N | 57 | BSW | BS |业务| |O | 0 | BOSS | BS |业务| |P | 38 | MSD | BS |业务| |Q | 38 | MSD | BS |业务| |R | 19 | BHV | BR |业务| |S | 0 | BCR | BR |业务| |T | 19 | BHV | BR |业务| |U | 2 | BSW | BS |业务| |V | 1 | BSW | BS |业务| |W | 57 | BSW | BS |业务| |X | 38 | MSD | BS |业务| |Y | 38 | MSD | BS |业务| +----+---------+----------+---------------+----------------+ 下面是添加的3列中的预期结果

LOB_名称2（这是LN现在产生的STF_的最大值）

57 BSW 0老板 112 MSD 42 BHV 公元前23年 55 BSW2 LifeCycleName2（这是LOB_Name2的最大汇总值之和）

224英国标准 65溴 业务领域2（这是LifeCycleName2汇总的总和）

289业务

---

您不能

sum

a

max

，因为如果您有相同的分组，那么它无论如何都是相同的金额。您可能需要有一个具有不同分组依据的内部和外部部件，例如：

select
  product_group,
  sum(max_cost)
from
(
  select 
    product,
    product_group,
    max(cost) as max_cost
  from
    orders
  group by
     product_group,product
) X 
group by product_group

---

这个假想的SQL将获取每个产品的最大成本，并将它们相加到产品组级别。这是我唯一能计算出你实际上需要

sum

a

max

的方法。你不能

sum

a

max

，因为不管怎样，如果你有相同的组别，它的金额是相同的。您可能需要有一个具有不同分组依据的内部和外部部件，例如：

select
  product_group,
  sum(max_cost)
from
(
  select 
    product,
    product_group,
    max(cost) as max_cost
  from
    orders
  group by
     product_group,product
) X 
group by product_group

---

这个假想的SQL将获取每个产品的最大成本，并将它们相加到产品组级别。这是我唯一能确定您实际上需要求和的方法。

sum

a

max

您需要添加另一个查询级别来求和您的最大值

想法是在一次选择中最大化，然后使用外部查询对结果求和。我在下面的例子中使用了AVG和MAX，但是，可以使用任何聚合函数

SELECT
    LocationID,
    MaxAverageSalePriceByLocation=MAX(AvgerageSalePriceByUserLocation)
FROM
(
    SELECT
        UserID,
        AvgerageSalePriceByUserLocation=AVG(SalePrice)
    FROM
        MyTable
    GROUP BY
      UserID,LocationID
)AS A 
GROUP BY
    LocationID

---

您需要添加另一个查询级别来求和最大值

想法是在一次选择中最大化，然后使用外部查询对结果求和。我在下面的例子中使用了AVG和MAX，但是，可以使用任何聚合函数

SELECT
    LocationID,
    MaxAverageSalePriceByLocation=MAX(AvgerageSalePriceByUserLocation)
FROM
(
    SELECT
        UserID,
        AvgerageSalePriceByUserLocation=AVG(SalePrice)
    FROM
        MyTable
    GROUP BY
      UserID,LocationID
)AS A 
GROUP BY
    LocationID

---

如果要求所有行的总和

Max

值，请使用

OVER（）

---

如果要对每个组的所有行

Max

值求和，请使用

OVER（Partition by）

---

注意将数字数据转换为浮点可能会导致近似值问题。。使用带精度和刻度的数值

如果要对所有行的

Max

值求和，请使用

OVER（）

---

如果要对每个组的所有行

Max

值求和，请使用

OVER（Partition by）

注意<

select
  product_group,
  sum(max_cost)
from
(
  select 
    product,
    product_group,
    max(cost) as max_cost
  from
    orders
  group by
     product_group,product
) X 
group by product_group

SELECT
    LocationID,
    MaxAverageSalePriceByLocation=MAX(AvgerageSalePriceByUserLocation)
FROM
(
    SELECT
        UserID,
        AvgerageSalePriceByUserLocation=AVG(SalePrice)
    FROM
        MyTable
    GROUP BY
      UserID,LocationID
)AS A 
GROUP BY
    LocationID

Sum(Max(CONVERT(FLOAT, Replace([STF_Now], 'N/A', '0'))))OVER() AS [STF2_Now]

Sum(Max(CONVERT(FLOAT, Replace([STF_Now], 'N/A', '0'))))OVER(partition by grp1,grp2,..)  AS [STF2_Now]

SELECT ... 
, SUM(AA2_Now_Max)
...
, SUM(STF2_Now_Max)
FROM(
SELECT  dbo.CCA_Merged.id, dbo.CCA_Merged.timeStamp, dbo.CCA_Merged.name, dbo.CCA_Merged.lN
    ,dbo.CCA_Merged.type, dbo.CCA_Merged.id2, dbo.CCA_Merged.aG 
    ,dbo.CCA_Merged.regionId, dbo.CCA_Merged.sgcc       
    ,convert(float,replace([SLC_Today],'N/A','0')) As [SLC_Today]
    ,convert(float,replace([AA_Now],'N/A','0')) As [AA_Now]
    ,MAX(convert(float,replace([AA_Now],'N/A','0'))) As [AA2_Now_Max]
    ,convert(float,replace([SLCO_Today],'N/A','0')) As [SLCO_Today]
    ,convert(float,replace([CABN_Today],'N/A','0')) As [CABN_Today]
    ,convert(float,replace([COF_Today],'N/A','0')) As [COF_Today]
    ,convert(float,replace([HT_Today],'N/A','0')) As [HT_Today]
    ,convert(float,replace(replace([CH_Today],'N/A','0'),'-','0')) As [CH_Today]
    ,convert(float,replace([SLC_Now],'N/A','0')) As [SLC_Now]
    ,convert(float,replace([SLCO_Now],'N/A','0')) As [SLCO_Now]
    ,convert(float,replace([SLC_Thirty],'N/A','0')) As [SLC_Thirty]
    ,convert(float,replace(replace([SLCO_Thirty],'N/A','0'),'-','0')) As [SLCO_Thirty]
    ,convert(float,replace([ACWT_Today],'N/A','0')) As [ACWT_Today]
    ,convert(float,replace([CQ_Now],'N/A','0')) As [CQ_Now]
    ,convert(float,replace([LCQ_Now],'N/A','0')) As [LCQ_Now]
    ,convert(float,replace([SLCH_Now],'N/A','0')) As [SLCH_Now]
    ,convert(float,replace([STF_Now],'N/A','0')) As [STF_Now]
    ,MAX(convert(float,replace([STF_Now],'N/A','0'))) As [STF2_Now_Max]
    ,dbo.Sheet1$.AreaOfBusiness, dbo.Sheet1$.LifeCycleName, dbo.Sheet1$.LOB_name    
FROM            dbo.Sheet1$ RIGHT OUTER JOIN
                 dbo.CCA_Merged ON dbo.Sheet1$.Skill_Name = dbo.CCA_Merged.lN   
Group by ROLLUP (stf_now) ,dbo.CCA_Merged.id, dbo.CCA_Merged.timeStamp, dbo.CCA_Merged.name, dbo.CCA_Merged.lN
        ,dbo.CCA_Merged.type, dbo.CCA_Merged.id2, dbo.CCA_Merged.aG ,dbo.CCA_Merged.regionId
        ,dbo.CCA_Merged.sgcc,AA_Now,SLC_Today,SLCO_Today,CABN_Today,COF_Today,HT_Today,CH_Today
        ,SLC_Now,SLCO_Now,SLC_Thirty,SLCO_Thirty,ACWT_Today,CQ_Now,LCQ_Now,SLCH_Now
        ,dbo.Sheet1$.AreaOfBusiness, dbo.Sheet1$.LifeCycleName, dbo.Sheet1$.LOB_name ) x
GROUP BY ...