Sql 如何按报告周创建事务表？

我对SQL非常陌生，我正在尝试通过报告周开始日期（星期一）来创建一个新表，并根据每个事务的报告周开始日期来匹配它们。如何在Google BigQuery SQL中做到这一点

我的桌子

Account_ID  | Order_ID |  Reporting_week_start_date
    001     | 1001     |       31 Dec 2018         
    001     | 1002     |       28 Jan 2019         
    001     | 1003     |       25 Feb 2019    
    002     | 1004     |       31 Dec 2018         
    002     | 1005     |       28 Jan 2019         
    002     | 1006     |       25 Feb 2019         

期望输出

    Reporting_week_start_date | Account_ID | Order_ID 
    31 Dec 2018               |     001    |  1001     #Transaction 1 (Account_ID 001)
    07 Jan 2019               |            |
    14 Jan 2019               |            |
    21 Jan 2019               |            |
    28 Jan 2019               |     001    |  1002     #Transaction 2 (Account_ID 001)
    04 Feb 2019               |            |
    11 Feb 2019               |            |
    18 Feb 2019               |            |
    25 Feb 2019               |     001    |  1003     #Transaction 3 (Account_ID 001)
    04 Mar 2019               |            |
    11 Mar 2019               |            |
    18 Mar 2019               |            |    

    31 Dec 2018               |     002    |  1004     #Transaction 1 (Account_ID 002)
    07 Jan 2019               |            |
    14 Jan 2019               |            |
    21 Jan 2019               |            |
    28 Jan 2019               |     002    |  1005     #Transaction 2 (Account_ID 002)
    04 Feb 2019               |            |
    11 Feb 2019               |            |
    18 Feb 2019               |            |
    25 Feb 2019               |     002    |  1006     #Transaction 3 (Account_ID 002)
    04 Mar 2019               |            |
    11 Mar 2019               |            |
    18 Mar 2019               |            |   

 ... to current_date()

这是你想要的吗

select wk, t.account_id, t.order_id
from unnest(generate_date_array(date('2018-12-31'), date('2019-03-18'), interval 1 week)) wk cross join
     (select distinct account_id
      from t
     ) a left join
     t
     on t.account_id = a.account_id and
        t.reporting_week_start_date = wk
order by a.account_id, wk;

---

如果您想对每个

账户ID

和

订单ID

的日期金额进行分组，我认为此查询可以帮助您获得所需的内容

为了测试查询，首先我创建了一个包含数据的新表，如您的表：

CREATE TABLE `project.dataset.reporting_week` ( 
    `Account_ID` STRING,
    `Order_ID` STRING,
    `Reporting_week_start_date` DATE);

INSERT INTO `project.dataset.reporting_week` VALUES
  ('001', '1001', PARSE_DATE('%Y/%m/%d', '2018/12/31')),
  ('001', '1001', PARSE_DATE('%Y/%m/%d', '2019/01/07')),
  ('001', '1001', PARSE_DATE('%Y/%m/%d', '2019/01/14')),
  ('001', '1001', PARSE_DATE('%Y/%m/%d', '2019/01/21')),
  ('001', '1002', PARSE_DATE('%Y/%m/%d', '2019/01/28')),
  ('001', '1002', PARSE_DATE('%Y/%m/%d', '2019/02/04')),
  ('001', '1002', PARSE_DATE('%Y/%m/%d', '2019/02/11')),
  ('001', '1002', PARSE_DATE('%Y/%m/%d', '2019/02/18')),
  ('001', '1003', PARSE_DATE('%Y/%m/%d', '2019/02/25')),
  ('001', '1003', PARSE_DATE('%Y/%m/%d', '2019/03/04')),
  ('001', '1003', PARSE_DATE('%Y/%m/%d', '2019/03/11')),
  ('001', '1003', PARSE_DATE('%Y/%m/%d', '2019/03/18')),
  ('002', '1004', PARSE_DATE('%Y/%m/%d', '2018/12/31')),
  ('002', '1004', PARSE_DATE('%Y/%m/%d', '2019/01/07')),
  ('002', '1004', PARSE_DATE('%Y/%m/%d', '2019/01/14')),
  ('002', '1004', PARSE_DATE('%Y/%m/%d', '2019/01/21')),
  ('002', '1005', PARSE_DATE('%Y/%m/%d', '2019/01/28')),
  ('002', '1005', PARSE_DATE('%Y/%m/%d', '2019/02/04')),
  ('002', '1005', PARSE_DATE('%Y/%m/%d', '2019/02/11')),
  ('002', '1005', PARSE_DATE('%Y/%m/%d', '2019/02/18')),
  ('002', '1006', PARSE_DATE('%Y/%m/%d', '2019/02/25')),
  ('002', '1006', PARSE_DATE('%Y/%m/%d', '2019/03/04')),
  ('002', '1006', PARSE_DATE('%Y/%m/%d', '2019/03/11')),
  ('002', '1006', PARSE_DATE('%Y/%m/%d', '2019/03/18'))

因此，我们可以运行查询并从上表创建新表

CREATE TABLE `project.dataset.reporting_week2` AS
  SELECT new_account, new_order, ARRAY(SELECT Reporting_week_start_date FROM 
  `project.dataset.reporting_week` WHERE Account_ID = new_account and Order_ID = 
   new_order ) as DATES
  FROM (SELECT distinct account_id as new_account, new_order FROM 
  `project.dataset.reporting_week`, (SELECT distinct order_id as new_order FROM 
  `project.dataset.reporting_week`)
  where order_id = new_order)

输出类似于：

---

@Junior coder，您将使用以下查询获得预期结果。我以21天为例。代替21天，您可以将日差设置为（CurrentDate-reporting\u start\u date）

输出。。

---

如果一周内有多个订单怎么办？

WITH
    DS AS (select Account_ID, Order_ID ,Reporting_week_start_date as start_date, DATE_ADD(Reporting_week_start_date, INTERVAL 21 DAY) as advanced_date,  from pay_recon_cl_dataset.reporting_week)
select GENERATE_DATE_ARRAY(start_date,advanced_date, interval 1 week),Account_ID,Order_ID from DS;