合并交易但保留所有人口统计列

Question

I have multiple transactions for the same people and i want to merge them all and get the total spent for each but i want to keep all of their demography variables in the same table, when i have tried some codes like SELECT DISTINCT it will just delete some duplicates instead of merging, my target is to put the customers into 2 groups Low and High value. Either someone who bought 1 item for $950 or who bought multiple items and spent more than $2500. (each transaction is just a single item no transaction has more than 1 item.

Here is my code so far. (i am preparing it for SAS)

Select        
CUS.FirstName        
,CUS.LastName        
,CUS.NumberChildrenAtHome        
,CUS.CommuteDistance        
,CUS.CustomerKey        
,FIS.SalesAmount        
,CUS.Gender        
,CUS.MaritalStatus        
,CUS.HouseOwnerFlag        
,CUS.NumberCarsOwned        
,CUS.YearlyIncome        
,CUS.TotalChildren        
,CUS.EnglishEducation AS Education        
,floor(DATEDIFF(DAY,BirthDate,getdate()))/365.25 AS AGE
,CASE        
WHEN FIS.UnitPrice >=950 OR FIS.SalesAmount >=2500 THEN 'High Value'        
ELSE 'Low Value'        
END AS 'Customer Value'        
From dbo.FactInternetSales AS FIS        
LEFT JOIN DBO.DimCustomer AS CUS        
ON FIS.CustomerKey = CUS.CustomerKey        
LEFT JOIN dbo.DimSalesTerritory AS DST        
ON FIS.SalesTerritoryKey = DST.SalesTerritoryKey

This is my first time using this type of question/answer site so sorry if i did something wrong.

Thank you

Answer 1

You should first SUM BY the amounts, then you can categorise based on the totals. The inner query will give you the sum for each person. The outer query will set the category based on your condition.

-- EDITED --

SELECT
    innerQuery.*,
    CASE WHEN Total_Unit_Price >= 950
    OR Total_Sales_Amount >= 2500 THEN 'High Value' ELSE 'Low Value' END AS 'Customer Value'
FROM
    (
        Select
            CUS.FirstName,
            CUS.LastName,
            CUS.NumberChildrenAtHome,
            CUS.CommuteDistance,
            CUS.CustomerKey,
            FIS.SalesAmount,
            CUS.Gender,
            CUS.MaritalStatus,
            CUS.HouseOwnerFlag,
            CUS.NumberCarsOwned,
            CUS.YearlyIncome,
            CUS.TotalChildren,
            CUS.EnglishEducation AS Education,
            floor(DATEDIFF(DAY, BirthDate, getdate())) / 365.25 AS AGE,
            SUM(FIS.UnitPrice) AS Total_Unit_Price,
            SUM(FIS.SalesAmount) AS Total_Sales_Amount
        From
            dbo.FactInternetSales AS FIS
            LEFT JOIN DBO.DimCustomer AS CUS ON FIS.CustomerKey = CUS.CustomerKey
            LEFT JOIN dbo.DimSalesTerritory AS DST ON FIS.SalesTerritoryKey = DST.SalesTerritoryKey
        GROUP BY
            CUS.FirstName,
            CUS.LastName,
            CUS.NumberChildrenAtHome,
            CUS.CommuteDistance,
            CUS.CustomerKey,
            FIS.SalesAmount,
            CUS.Gender,
            CUS.MaritalStatus,
            CUS.HouseOwnerFlag,
            CUS.NumberCarsOwned,
            CUS.YearlyIncome,
            CUS.TotalChildren,
            CUS.EnglishEducation,
            floor(DATEDIFF(DAY, BirthDate, getdate())) / 365.25
    ) innerQuery