如何僅對 postgres 中的唯一值求和？

Question

我在下面有一個表格，我試圖通過唯一的訂單號對order_value分組求和。 但是，我嘗試過的查詢返回了全部金額。 預期回報為： 34.24但實際回報為50.37 。 如何獲得正確的返回值？

訂單號	訂單價值	ID
#1005	16.03	1
#1005	16.03	2
#1006	18.21	3

我的查詢：

SELECT order_number, 
       SUM(order_value) 
FROM customer_response 
GROUP BY order_number ;

我也試過：

SELECT SUM(order_value) 
FROM customer_response 
GROUP BY order_number, order_value

SELECT SUM(order_value) 
FROM customer_response 
GROUP BY order_number

這似乎應該很簡單，但我只是不確定我哪里出錯了。

Answer 1

與您的查詢非常相似，但首先只提取order_number的不同行。

SELECT order_number, SUM(order_value)
FROM
(
 select distinct on (order_number) *
 from customer_response
) t 
GROUP BY order_number;

如果您需要獲取所有訂單的總和，請刪除分組。

SELECT SUM(order_value)
FROM
(
 select distinct on (order_number) *
 from customer_response
) t;

SQL 小提琴

編輯
實際上，由於每個order_number求和只剩下一行，因此分組是沒有意義的。 所以第一個查詢變得簡單

 select distinct on (order_number) *
 from customer_response;

Answer 2

你可以試試這個：

SELECT sum(t.order_avg)
  FROM
     ( SELECT avg(order_value) AS order_avg
         FROM customer_response
        GROUP BY order_number
     ) AS t

Answer 3

如果要獲取每個Order_Number的所有不同Order_Value的總和，可以按Order_Number以獲取每個組中的總和，然后使用SUM() window function 獲得總數：

SELECT DISTINCT SUM(SUM(DISTINCT Order_Value)) OVER () total_value
FROM customer_response t
GROUP BY Order_Number;

請參閱演示。