动态日期范围内的计数

时间:2019-02-28 18:10:21

标签: sql google-bigquery

我有一个表,其中包含以下字段:customer_idstart_trial_dateend_trial_date。我正在尝试编写一个查询,可以用来计算给定日期的customer_id

|+-------------+------------+------------+
| customer_id | start_date |  end_date  |
+-------------+------------+------------+
|           1 | 2017-02-03 | 2017-05-01 |
|           2 | 2017-04-07 | 2017-09-01 |
|           3 | 2017-03-02 | 2018-03-04 |
|           4 | 2013-02-25 | 2015-01-22 |
|           5 | 2015-11-10 | 2016-03-25 |
|     ....    |     ....   |     ....   |
+-------------+------------+------------+

我将如何编写查询以生成一个结果集,该结果集的所有日期都在一个计数范围内,其中customer_id的有效期限包括该日期?

预期输出:

+------------+-----------+
|    date    | customers |
+------------+-----------+
| 2013-01-01 | 0         |
| ….         | ….        |
| 2017-04-20 | 3         |
| …..        | ….        |
| 2018-12-31 | ….        |
+------------+-----------+

如果这很重要,我正在使用BigQuery。我曾考虑过创建一个助手表,该表列出一个范围的所有日期,然后尝试将其加入到我的表中并进行计数,但是在这种方法中我没有很好的联接键。

2 个答案:

答案 0 :(得分:1)

创建日历表是一个很好的起点。一旦有了该表(例如表calendar的列为calendar_date),就可以使用LEFT JOIN和聚合了:

SELECT c.calendar_date, COUNT(t.customer_id) customers
FROM calendar c
LEFT JOIN mytable t 
    ON c.calendar_date >= t.start_date AND c.calendar_date <= t.end_date
GROUP BY c.calendar_date

注意:您可能需要根据您的确切要求调整不平等条件(>=><=<)。

答案 1 :(得分:1)

以下是用于BigQuery标准SQL

#standardSQL
WITH calendar AS (
  SELECT day
  FROM (
    SELECT MIN(start_date) min_date, MAX(end_date) max_date
    FROM `project.dataset.table`
  ), UNNEST(GENERATE_DATE_ARRAY(min_date, max_date)) day
)
SELECT day, COUNTIF(day BETWEEN start_date AND end_date) customers
FROM calendar, `project.dataset.table`
GROUP BY day

您可以使用下面的示例中的虚拟数据来测试,玩游戏

#standardSQL
WITH `project.dataset.table` AS (
  SELECT 1 customer_id, DATE '2017-01-01' start_date, DATE '2017-01-05' end_date UNION ALL
  SELECT 2, '2017-01-03', '2017-01-04' UNION ALL
  SELECT 3, '2017-01-04', '2017-01-06' UNION ALL
  SELECT 4, '2017-01-10', '2017-01-12' UNION ALL
  SELECT 5, '2017-01-12', '2017-01-13' 
), calendar AS (
  SELECT day
  FROM (
    SELECT MIN(start_date) min_date, MAX(end_date) max_date
    FROM `project.dataset.table`
  ), UNNEST(GENERATE_DATE_ARRAY(min_date, max_date)) day
)
SELECT day, COUNTIF(day BETWEEN start_date AND end_date) customers
FROM calendar, `project.dataset.table`
GROUP BY day
-- ORDER BY day   

有结果

Row day         customers    
1   2017-01-01  1    
2   2017-01-02  1    
3   2017-01-03  2    
4   2017-01-04  3    
5   2017-01-05  2    
6   2017-01-06  1    
7   2017-01-07  0    
8   2017-01-08  0    
9   2017-01-09  0    
10  2017-01-10  1    
11  2017-01-11  1    
12  2017-01-12  2    
13  2017-01-13  1