需要帮助为没有订单的月份生成 NULL 条目

Question

我必须创建显示 table (Tbl) 中所有字段的输出，并创建一个额外的列来按月计算每个客户的累计总和（例如，如果客户在 4 月份有两次销售，新列将包含这些销售额和两行中任何先前销售额的总和）。我能做的就这么多了。

我的问题是每月为每个客户生成行即使他们没有销售额，并且累计列仍然正确显示上个月的累计总和。

期望的输出： Picture Link

Customer_ID Order_ID    Order_Date  Order_Amt_Total_USD Month_ID    Cum_Total_By_Month
John        123        4/4/2019      30                  Jun-19          120
John        124        4/12/2019     90                  Jun-19          120
Mark       null         null        null                 Jun-19           0
Sally       150        4/20/2019     50                  Jun-19           50
John        null         null       null                 Jul-19           120
Mark        165        7/7/2019      80                  Jul-19           170
Mark        166        7/7/2019      90                  Jul-19           170
Sally       160        7/5/2019      75                  Jul-19           125
John        null        null         null                Aug-19           120
Mark        null        null         null                Aug-19           170
Sally       null        null         null                Aug-19           125

我将在下面列出代码，但这是一个 link 到 SQL fiddle 的示例数据和我正在处理的片段的两个查询（与来自本网站上优秀人士的帮助）。 http://sqlfiddle.com/#!15/1d86b/11

我可以使用第一个查询按客户和月份生成所需的累计运行总和。

我还可以生成一个基数 table，它在第二个查询中为每个客户每个月提供一个 month_id。

我需要帮助将这两者结合起来，以便在 months/Customers 没有任何销售时生成包含空行的所需输出。

有什么想法吗？谢谢！

-- Generates cumulative total by month by Customer, but only shows when they have a sale
SELECT 
    Customer_ID, Order_Date, order_id, Order_Amt_Total_USD,
    to_char(date_trunc('month', Order_Date), 'Mon YYYY') AS mon_text,
    (Select 
     sum(Order_Amt_Total_USD) 
        FROM tbl t2
         WHERE t2.Customer_ID = t.Customer_ID
         AND date_trunc('month', t2.Order_Date) <= t.Order_Date ) AS Cumulative
FROM    tbl t
GROUP BY mon_text, Customer_ID, Order_Date, order_id, Order_Amt_Total_USD
ORDER BY date_trunc('month', Order_Date), Customer_ID, Order_Date
;

-- Generates Proper List of All Month IDs for each Customer from entered date through today
WITH temp AS (
   SELECT date_trunc('month', Order_Date) AS mon_id
  FROM tbl
  )
Select
Customer_ID,
to_char(mon_id, 'Mon YYYY') AS mon_text
From tbl,
generate_series('2015-01-01'::date, now(), interval '1 month') mon_id
LEFT  JOIN temp USING (mon_id)
GROUP BY mon_id,Customer_ID
;

Answer 1

下面的示例显示了如何使用分区依据来实现输出-

架构

CREATE TABLE tbl (Order_ID varchar(10), Customer_ID varchar(10), Order_Date date, Order_Amt_Total_USD bigint);
    
INSERT INTO tbl (Order_ID, Customer_ID, Order_Date, Order_Amt_Total_USD)
VALUES
    ('100', 'qwe', '2015-08-04', 6),
    ('101', 'qwe', '2015-05-20', 7),
    ('102', 'qwe', '2015-04-08', 8),
    ('103', 'qwe', '2015-04-07', 9),
    ('109', 'aaa', '2015-04-28', 1),
    ('110', 'aaa', '2015-04-28', 2),
    ('111', 'aaa', '2015-05-19', 3),
    ('112', 'aaa', '2015-08-06', 4),
    ('113', 'aaa', '2015-08-27', 5),
    ('114', 'aaa', '2015-08-07', 6)

查询

select  Order_Date , Customer_ID, Order_Amt_Total_USD,
       sum(Order_Amt_Total_USD) over (partition by Customer_ID order by Order_Date) as cumulative
from tbl;

Answer 2

根据您的描述，您可以将 window 函数与 generate_series() 结合使用：

SELECT c.Customer_ID, mon,
       SUM(Order_Amt_Total_USD) as month_total,
       SUM(SUM(Order_Amt_Total_USD)) OVER (PARTITION BY c.Customer_ID ORDER BY mon) as running_total
FROM (SELECT DISTINCT Customer_Id FROM tbl) c CROSS JOIN
     generate_series('2015-01-01'::date, now(), interval '1 month') mon LEFT JOIN
     tbl t
     ON t.Customer_Id = c.customer_id and
        date_trunc('month', t.Order_Date) = mon
GROUP BY c.Customer_ID, mon
ORDER BY 1, 2;

Here 是 SQL Fiddle.

需要帮助为没有订单的月份生成 NULL 条目

Need Help Generating NULL Entries for Months with No Orders

sql

null

cumulative-sum