Sql server 2008 如何按频率订购而不进行聚合？

概述：
比较不同账户代理农产品的日常交易，找出代理和公司持有量之间的差异，仅显示新情况；当前交易中的代码，但不是以前交易中的代码

目标：
按最高频率的代码（即苹果）对交易进行排序

当前代码：

SELECT 
    Count(CurrentTrade.Code) AS CountOfCode, 
    CurrentTrade.portfolio_code, 
    CurrentTrade.Type, 
    CurrentTrade.Code, 
    sum(CurrentTrade.Custodian_holding) AS AgentHolding, 
    sum(CurrentTrade.Company_holding) AS CompanyHolding, 
    sum(CurrentTrade.Difference) AS TotalDifference
FROM 
    CurrentTrade 
    LEFT JOIN PrevTrade 
    ON CurrentTrade.Company_holding = PrevTrade.Company_holding
WHERE 
    PrevTrade.Company_holding) Is Null
GROUP BY CurrentTrade.Code, CurrentTrade.portfolio_code, CurrentTrade.Code, CurrentTrade.Type
ORDER BY count(CurrentTrade.Code);

目前的结果：

 CountOfCode        account_code        Type        Code        AgentHodling        CompanyHolding        TotalDifference
 1                  acc001049        prod        mango        2500                  0        2500
 1                  acc636903        prod        mango        481081.29        481081.285        0.01
 1                  acc179186        prod        mango        453639.61        511285.81        -57646.2
 1                  acc179185        prod        mango        2775               2650        125
 1                  acc170001        prod        grapes       0                     1500        -1500
 1                  acc170000        prod        grapes       2500        0        2500
 1                  acc107104        prod        apple        49797.22        49797.28        -0.06
 1                  acc107103        prod        apple        0        -0.06        0.06
 1                  acc107102        prod        apple        2000        500        1500
 1                  acc107101        prod        apple        0        1500        -1500
 1                  acc001189        prod        apple        0        1090.513        -1090.513

参考：

我收到了这个（）的答案，所以这是一个跟进

问题:

---

是否可以在不汇总所有结果的情况下按频率排序，并且仍然按最高频率（苹果、芒果、葡萄）显示订单，包括相关的账号
如果对order by使用sum，“不能在order by中使用聚合函数”，因此我不确定我是否可以这样做，并且仍然保持帐号分开？

将

order by

子句拉出到外部查询中，内部查询将计算频率，如下所示：

SELECT * FROM (
SELECT 
    Count(CurrentTrade.Code) AS CountOfCode, 
    CurrentTrade.portfolio_code, 
    CurrentTrade.Type, 
    CurrentTrade.Code, 
    sum(CurrentTrade.Custodian_holding) AS AgentHolding, 
    sum(CurrentTrade.Company_holding) AS CompanyHolding, 
    sum(CurrentTrade.Difference) AS TotalDifference
FROM 
    CurrentTrade 
    LEFT JOIN PrevTrade 
    ON CurrentTrade.Company_holding = PrevTrade.Company_holding
WHERE 
    PrevTrade.Company_holding) Is Null
GROUP BY CurrentTrade.Code, CurrentTrade.portfolio_code, CurrentTrade.Code, CurrentTrade.Type
) ORDER BY CountOfCode ASC

---

谢谢你的解决方案！我最终创建了两个报告：1）没有组合代码作为摘要，按计数顺序正确；2）有组合代码作为详细信息。