如何计算同一列但不同类别的日期之间的差异？答案

【问题标题】：How can I calculate difference between dates in the same column but different categories?如何计算同一列但不同类别的日期之间的差异？
【发布时间】：2016-12-17 08:42:43
【问题描述】：

我相信有人会指出这是一个基本的重塑问题，但我想我在谷歌搜索方面很差，所以我将我的问题提供给 Stack Overflow 的诸神。

我的问题是我想计算来自同一列但在另一列（排名）中具有不同值的两个日期（排名 1 和 2）之间的天数。这是数据集架构的示例：

User   Date        Rank
Bob    2016-12-01    3
Bob    2016-12-07    2
Bob    2016-12-10    1

我想要的是这样的：

User   Date1        Date2        DaysBetween
Bob    2016-12-07   2016-12-10        3

理想情况下，我想在 Tableau 中执行此操作，但 SQL/BigQuery 也可以。

【问题讨论】：

一个用户是否曾经有不止一行具有相同的排名？你是完全无视3级吗？或者你也想要排名 3 和 2 之间的天数？在您的示例中，排名似乎只是日期的相反顺序。总是这样吗？
对于 Tableau 解决方案，我可能会使用表格计算。取决于以上问题的答案。

标签： sql google-bigquery tableau-api

【解决方案1】：

下面试试

WITH YourTable AS (
  SELECT 'Bob' AS User, DATE '2016-12-01' AS Date, 3 AS Rank UNION ALL
  SELECT 'Bob' AS User, DATE '2016-12-07' AS Date, 2 AS Rank UNION ALL
  SELECT 'Bob' AS User, DATE '2016-12-10' AS Date, 1 AS Rank 
)
SELECT 
  User, 
  MAX(CASE WHEN Rank = 2 THEN Date END) AS Date1,
  MAX(CASE WHEN Rank = 1 THEN Date END) AS Date2,
  DATE_DIFF(MAX(CASE WHEN Rank = 1 THEN Date END), 
            MAX(CASE WHEN Rank = 2 THEN Date END), DAY) AS DaysBetween
FROM YourTable
GROUP BY User

注意 - 这是在 BigQuery Standard SQL 中

【讨论】：

【解决方案2】：

另一种选择

WITH YourTable AS (
  SELECT 'Bob' AS User, DATE '2016-12-01' AS Date, 3 AS Rank UNION ALL
  SELECT 'Bob' AS User, DATE '2016-12-07' AS Date, 2 AS Rank UNION ALL
  SELECT 'Bob' AS User, DATE '2016-12-10' AS Date, 1 AS Rank 
)
SELECT 
  User, Date1, Date2, 
  DATE_DIFF(Date2, Date1, DAY) AS DaysBetween
FROM (
  SELECT 
    User, Rank, Date as Date2, 
    LEAD(Date) OVER(PARTITION BY User ORDER BY Date DESC) AS Date1
  FROM YourTable
)
WHERE Rank = 1

【讨论】：