如何在不使用 Union all with SQL 的情况下添加两个具有最旧和最新日期的新列

Question

我正在尝试在 BigQuery 中创建一个新的 table，其中包括具有最新和最旧日期的两列。

来源Table:

date	concatenatedString
27/08/2019	Housesystems-ActivationHouses-Multiple-Display & Video
03/09/2019	Housesystems-ActivationHouses-Multiple-Display & Video
28/11/2019	Housesystems-ActivationHouses-Multiple-Display & Video
01/09/2019	Housesystems-ActivationHouses-Multiple-Paid Social
11/09/2020	Housesystems-ActivationHouses-Multiple-Paid Social
06/07/2021	Housesystems-ActivationHouses-Multiple-Paid Social

结果：

concatenatedString	Oldest_Date	Newest_Date
Housesystems-ActivationHouses-Multiple-Paid Social	01/09/2019	06/07/2021
Housesystems-ActivationHouses-Multiple-Display & Video	27/08/2019	28/11/2019

这是我的查询：

        SELECT
      concatenatedString,
      MIN(FirstLastDate) AS KampagneStart,
      MAX(FirstLastDate) AS KampagneEnd
    FROM (
      SELECT
        MAX(date) AS FirstLastDate,
        CONCAT(ifnull(Campaign_Name,
            ''),'|', ifnull(Campaign_Category,
            ''),'|',ifnull(Product_Category,
            ''),'|',ifnull(Powerbrand_Powerline,
            ''),'|',ifnull(Channel,
            '')) AS concatenatedString
      FROM
        `bigquery-project.ETL_Ready.Union`
      GROUP BY
        concatenatedString
      UNION ALL
      SELECT
        MIN(date) AS FirstLastDate,
        CONCAT(ifnull(Campaign_Name,
            ''),'|', ifnull(Campaign_Category,
            ''),'|',ifnull(Product_Category,
            ''),'|',ifnull(Powerbrand_Powerline,
            ''),'|',ifnull(Channel,
            '')) AS concatenatedString
      FROM
        `bigquery-project.ETL_Ready.Union`
      GROUP BY
        concatenatedString )
    GROUP BY
      concatenatedString

问题是在源代码 table 中，我在单独的列中获取了 concatenatedString 的值，有时我会聚合另一个值。这就是为什么这个查询对我来说不是很灵活。

有谁知道不使用 UNION ALL 就能达到相同结果的有效方法吗？

Answer 1

考虑以下方法

select concatenatedString, 
  min(date) as Oldest_Date,
  max(date) as Newest_Date
from your_table
group by concatenatedString

如果应用于您问题中的示例数据 - 输出为

另一种选择是

select * from your_table
pivot (min(date) as Oldest, max(date) as Newest for 'Date' in ('Date'))

具有相同的输出

P.S。下面是我用于“测试”的 CTE

with your_table as (
  select '2019-08-27' date, 'Housesystems-ActivationHouses-Multiple-Display & Video' concatenatedString union all
  select '2019-09-03', 'Housesystems-ActivationHouses-Multiple-Display & Video' union all
  select '2019-11-28', 'Housesystems-ActivationHouses-Multiple-Display & Video' union all
  select '2019-09-01', 'Housesystems-ActivationHouses-Multiple-Paid Social' union all
  select '2020-09-11', 'Housesystems-ActivationHouses-Multiple-Paid Social' union all
  select '2021-07-06', 'Housesystems-ActivationHouses-Multiple-Paid Social' 
)

如何在不使用 Union all with SQL 的情况下添加两个具有最旧和最新日期的新列

How to add two new Columns with the oldest and newest date without using Union all with SQL in BigQuery

sql

google-bigquery