在 BigQuery 中提取当月的工作日和工作日

Question

我想做两件事 - 我正在尝试 a.) 仅提取一个月内的工作日和 b.) 它所在月份的工作日。

所以我知道工作日部分可以通过使用 DAYOFWEEK 函数并去掉等于 1 或 7 的所有天数来完成，但我对如何解决每月的工作日部分感到有些困惑。例如，今天是本月的第 21 个工作日，但另一个月的第 21 个工作日可能不是 29 日，所以这种比较没有意义。这可能是一个周末。为了简单起见，我们不必担心在典型工作日（即阵亡将士纪念日）的假期。

所以我的目标基本上是有一个输出给你 -

第 1 列：ID（可以说这是一些随机生成的三个字符的字母数字 ID）

第 2 列：开始日期

第 3 列：星期几

第 4 列：start_date 月内的工作日

有人有什么建议吗？

Answer 1

提取开始日期和结束日期之间的工作日

在下面的示例中，开始日期为“2016-06-01”，结束日期为“2016-06-30”

SELECT 
  DATE(DATE_ADD(TIMESTAMP('2016-06-01'), pos - 1, "DAY")) AS business_day
FROM (
     SELECT ROW_NUMBER() OVER() AS pos, *
     FROM (FLATTEN((
     SELECT SPLIT(RPAD('', 1 + DATEDIFF(TIMESTAMP('2016-06-30'), TIMESTAMP('2016-06-01')), '.'),'') AS h
     FROM (SELECT NULL)),h
)))
HAVING DAYOFWEEK(business_day) NOT IN (1,7)

查找开始日期和结束日期之间的第 N 个工作日

在下面的示例中，开始日期为“2016-06-01”，结束日期为“2016-06-30”，N = 21

SELECT 
  business_day
FROM (
  SELECT
    business_day,
    ROW_NUMBER() OVER(ORDER BY business_day) AS number
  FROM (
    SELECT 
      DATE(DATE_ADD(TIMESTAMP('2016-06-01'), pos - 1, "DAY")) AS business_day
    FROM (
         SELECT ROW_NUMBER() OVER() AS pos, *
         FROM (FLATTEN((
         SELECT SPLIT(RPAD('', 1 + DATEDIFF(TIMESTAMP('2016-06-30'), TIMESTAMP('2016-06-01')), '.'),'') AS h
         FROM (SELECT NULL)),h
    )))
    HAVING DAYOFWEEK(business_day) NOT IN (1,7)
  )
)
WHERE number = 21

Answer 2

下面试试

SELECT
  start_date, 
  business_day, 
  DAYOFWEEK(business_day) AS day_of_week,
  ROW_NUMBER() OVER(PARTITION BY start_date ORDER BY business_day) AS business_day_number
FROM (
  SELECT 
    start_date, DATE(DATE_ADD(TIMESTAMP(start_date), pos - 1, "DAY")) AS business_day
  FROM (
    SELECT start_date, ROW_NUMBER() OVER() AS pos
    FROM (FLATTEN((
      SELECT start_date, SPLIT(RPAD('', 1 + DATEDIFF(TIMESTAMP(end_date), TIMESTAMP(start_date)), '.'),'') AS h
      FROM 
        (SELECT '2016-01-01' AS start_date, '2016-01-31' AS end_date),
        (SELECT '2016-03-01' AS start_date, '2016-03-31' AS end_date),
        (SELECT '2016-06-01' AS start_date, '2016-06-30' AS end_date)
    ), h
  )))
  HAVING DAYOFWEEK(business_day) NOT IN (1,7)

)

在我看来 - 你改变了原来的问题 - 所以我觉得提供单独的答案而不是添加到我原来的答案是合适的

表名会去哪里？

假设 YourTable 如下所示：

start_date  end_date     
2016-01-01  2016-01-31   
2016-03-01  2016-03-31   
2016-06-01  2016-06-30

查询看起来像

SELECT
  start_date, 
  business_day, 
  DAYOFWEEK(business_day) AS day_of_week,
  ROW_NUMBER() OVER(PARTITION BY start_date ORDER BY business_day) AS business_day_number
FROM (
  SELECT 
    start_date, DATE(DATE_ADD(TIMESTAMP(start_date), pos - 1, "DAY")) AS business_day
  FROM (
    SELECT start_date, ROW_NUMBER() OVER() AS pos
    FROM (FLATTEN((
      SELECT start_date, SPLIT(RPAD('', 1 + DATEDIFF(TIMESTAMP(end_date), TIMESTAMP(start_date)), '.'),'') AS h
      FROM YourTable), h
  )))
  HAVING DAYOFWEEK(business_day) NOT IN (1,7)

)