如何实现基于单条记录的多条记录输出答案

【问题标题】：How to achieve multiple records output with based on single record如何实现基于单条记录的多条记录输出
【发布时间】：2019-03-13 23:17:21
【问题描述】：

我有一个“经理”表，其中包含如下数据

我期待如下输出

或我期望的另一种输出格式是

条件是经理 1001 于 2018 年加入，结束日期为 9999，因此他在 2018 年、2019 年和 2020 年活跃

经理1004于2018年加入，同年离开公司，所以只在2018年活跃

请帮助我如何实现这一目标

【问题讨论】：

为什么要在单独的行中使用它？想要的结果没有任何意义。
是的，我希望它们在单独的行中，这就是我面临的问题，如果我得到第二个输出格式也可以

标签： sql google-bigquery

【解决方案1】：

用它建立一个年份列表和JOIN：

SELECT manager_id, yearnum, 'Active' AS status
FROM UNNEST(GENERATE_ARRAY(2018, 2020)) AS yearnum
JOIN managers ON yearnum BETWEEN EXTRACT(year FROM eff_start_date)
                             AND EXTRACT(year FROM eff_end_date)

【讨论】：

【解决方案2】：

以下是 BigQuery 标准 SQL

#standardSQL
WITH years AS (
  SELECT EXTRACT(YEAR FROM year) year 
  FROM ( SELECT 
    (SELECT MIN(eff_start_date) FROM `project.dataset.managers`) AS min_date, 
    (SELECT MAX(eff_end_date) FROM `project.dataset.managers` WHERE eff_end_date != '9999-12-31') max_date  
  ), UNNEST(GENERATE_DATE_ARRAY(DATE_TRUNC(min_date, YEAR), DATE_TRUNC(max_date, YEAR), INTERVAL 1 YEAR)) year
), managers_list AS (
  SELECT manager_id, status, EXTRACT(YEAR FROM eff_start_date) start_year, EXTRACT(YEAR FROM eff_end_date) end_year
  FROM `project.dataset.managers`
)
SELECT manager_id, year, status 
FROM years y, managers_list m
WHERE year BETWEEN start_year AND end_year

您可以使用您问题中的示例数据进行测试，使用上面的示例，如下例所示

#standardSQL
WITH `project.dataset.managers` AS (
  SELECT 1001 manager_id, 'Active' status, DATE '2018-02-10' eff_start_date, DATE '9999-12-31' eff_end_date UNION ALL
  SELECT 1002, 'Active', '2018-02-14', '2020-12-31' UNION ALL
  SELECT 1003, 'Active', '2018-02-16', '2019-02-15' UNION ALL
  SELECT 1004, 'Active', '2018-02-16', '2018-12-31' 
), years AS (
  SELECT EXTRACT(YEAR FROM year) year 
  FROM ( SELECT 
    (SELECT MIN(eff_start_date) FROM `project.dataset.managers`) AS min_date, 
    (SELECT MAX(eff_end_date) FROM `project.dataset.managers` WHERE eff_end_date != '9999-12-31') max_date  
  ), UNNEST(GENERATE_DATE_ARRAY(DATE_TRUNC(min_date, YEAR), DATE_TRUNC(max_date, YEAR), INTERVAL 1 YEAR)) year
), managers_list AS (
  SELECT manager_id, status, EXTRACT(YEAR FROM eff_start_date) start_year, EXTRACT(YEAR FROM eff_end_date) end_year
  FROM `project.dataset.managers`
)
SELECT manager_id, year, status 
FROM years y, managers_list m
WHERE year BETWEEN start_year AND end_year 
-- ORDER BY manager_id, year

结果

Row manager_id  year    status   
1   1001        2018    Active   
2   1001        2019    Active   
3   1001        2020    Active   
4   1002        2018    Active   
5   1002        2019    Active   
6   1002        2020    Active   
7   1003        2018    Active   
8   1003        2019    Active   
9   1004        2018    Active

【讨论】：