對一系列數據的GROUP By子句SQL

Question

我正在編寫報告，並且在正確返回一系列數據（數據倉庫可能是理想的，但現在不是一個選項）時遇到了一些麻煩。實際上，我需要加入並報告2個表格...

交易和收據。交易包含收費金額和收據包含已付金額。我的報告需要顯示：

LastName | FirstName | Company | Location | Total | 30 | 60 | 90 

--------------------------------------------------------------------- 
Tom  | Clark | Microsoft | Washington | $300 | $80 | $100 | $120 

30,60,90哪裏有水桶顯示所欠金額在30天前，60天等，這是我掙扎。我可以毫無問題地獲得其他值。這是我迄今：

select 
    st.Client_Name_Last, 
    st.Client_Name_First, 
    st.Location_Company, 
    st.Location_Address_City, 
    sum((st.Billing_AmountPerUnitAllowed * st.Billing_NumberOfUnits) - coalesce(r.PaymentAmount, 0)) as Total, 
    (select sum((st.Billing_AmountPerUnitAllowed * st.Billing_NumberOfUnits) - coalesce(r.PaymentAmount, 0)) 
     where DateDiff(day, st.service_date, @effectiveDate) > 0 and DateDiff(day, st.service_date, @effectiveDate) < 30) as '30', 
    (select sum((st.Billing_AmountPerUnitAllowed * st.Billing_NumberOfUnits) - coalesce(r.PaymentAmount, 0)) 
     where DateDiff(day, st.service_date, @effectiveDate) >= 30 and DateDiff(day, st.service_date, @effectiveDate) < 60) as '60' 

from 
    ServiceTransactions st 
    join Claims c on st.Claim_Id = c.Id 
    left outer join Receipts r on c.Id = r.ClaimId 

group by 
    st.Client_Name_Last,  
    st.Client_Name_First, 
    st.Location_Company, 
    st.Location_Address_City 

這當然不行，因爲st.Service_Date是在頂級SELECT語句，這會導致一個錯誤，因爲它不是在聚合或group by子句。我考慮過使用Common Table Expression，但不知道如何最好地利用它。任何洞察力將是最讚賞。

謝謝你的時間！

Answer 1

你想要條件聚合。這使case內的sum()：使用

sum(case when DateDiff(day, st.service_date, @effectiveDate) > 0 and DateDiff(day, st.service_date, @effectiveDate) < 30) 
     then (st.Billing_AmountPerUnitAllowed * st.Billing_NumberOfUnits) - 
       coalesce(r.PaymentAmount, 0) 
      else 0 
    end) as days_30, 
sum(case when DateDiff(day, st.service_date, @effectiveDate) >= 30 and DateDiff(day, st.service_date, @effectiveDate) < 60) 
     then (st.Billing_AmountPerUnitAllowed * st.Billing_NumberOfUnits) - 
       coalesce(r.PaymentAmount, 0) 
      else 0 
    end) as days_60