合并 2 个 SQL 查询的结果答案

【问题标题】：Combine results from 2 SQL queries合并 2 个 SQL 查询的结果
【发布时间】：2020-04-06 12:00:02
【问题描述】：

我想比较两个具有相似数据的表。以下查询提供了每个表中的员工数量：

SELECT Extract_Month, COUNT(Employee_Number) as Staff
FROM Staff
GROUP BY Extract_Month;

SELECT Extract_Month, COUNT(Employee_Number) as MX
FROM Staff_mx
GROUP BY Extract_Month;

结果如下

2019-04-01  17246
2019-05-01  17194
2019-06-01  17252

2019-04-01  17140
2019-05-01  17194
2019-06-01  17434

我想要一个查询以以下格式组合结果：

            Staff MX
2019-04-01  17246 17140
2019-05-01  17194 17194
2019-06-01  17252 17434

我试过这个，其中包含我想要的结果，但没有正确呈现：

SELECT s.Extract_Month, COUNT(s.Employee_Number) as Staff, COUNT(mx.Employee_Number) as MX
FROM Staff s
    LEFT JOIN
        staff_mx mx
        ON s.Employee_Number = mx.Employee_Number 
        AND s.Extract_Month = mx.Extract_Month 
GROUP BY s.Extract_Month
UNION ALL
SELECT mx.Extract_Month, COUNT(s.Employee_Number) as Staff, COUNT(mx.Employee_Number) as MX
FROM Staff s
    RIGHT JOIN
        staff_mx mx
        ON s.Employee_Number = mx.Employee_Number 
        AND s.Extract_Month = mx.Extract_Month 
GROUP BY mx.Extract_Month

这给了我以下，但不是预期的结果：

            Staff   MX    
2019-04-01  17246   16830
2019-05-01  17194   16820
2019-06-01  17252   16784
2019-04-01  16830   17140
2019-05-01  16820   17194
2019-06-01  16784   17434

任何帮助将不胜感激！

【问题讨论】：

标签： sql sql-server join union

【解决方案1】：

你可以使用JOIN:

SELECT s.Extract_Month, s.staff, sm.mx
FROM (SELECT Extract_Month, COUNT(Employee_Number) as Staff
      FROM Staff
      GROUP BY Extract_Month
     ) s JOIN
     (SELECT Extract_Month, COUNT(Employee_Number) as MX
      FROM Staff_mx
      GROUP BY Extract_Month
     ) sm
     ON sm.Extract_Month = s.Extract_Month;

如果两个表的月份不一定相同，则使用FULL JOIN 和COALESCE()：

SELECT COALESCE(s.Extract_Month, sm.Extract_Month) as Extract_Month,
       COALESCE(s.staff, 0) as staff,
       COALESCE(sm.mx, 0) as mx
FROM (SELECT Extract_Month, COUNT(Employee_Number) as Staff
      FROM Staff
      GROUP BY Extract_Month
     ) s FULL JOIN
     (SELECT Extract_Month, COUNT(Employee_Number) as MX
      FROM Staff_mx
      GROUP BY Extract_Month
     ) sm
     ON sm.Extract_Month = s.Extract_Month;

【讨论】：

完美！像魅力一样工作，谢谢。 JOIN是一种什么样的join？
一个INNER JOIN@PhilCollins。

【解决方案2】：

做条件聚合：

with cte as (
     SELECT Extract_Month, Employee_Number as Staff, 'staff' AS flag
     FROM Staff
     UNION ALL
     SELECT Extract_Month, Employee_Number as Staff, 'mx' AS flag
     FROM Staff_mx
)
SELECT Extract_Month, 
       COUNT(CASE WHEN flag = 'Staff' THEN Employee_Number END) AS Staff,
       COUNT(CASE WHEN flag = 'mx' THEN Employee_Number END) AS mx
FROM cte c
GROUP BY Extract_Month;

【讨论】：

【解决方案3】：

也许你想要的是这样的：

SELECT ISNULL(S.Extract_Month,Sm.Extract_Month) AS Extract_Month,
       S.Staff,
       Sm.MX
FROM (SELECT Extract_Month, COUNT(Employee_Number) as Staff
      FROM Staff
      GROUP BY Extract_Month) S
     FULL OUTER JOIN (SELECT Extract_Month, COUNT(Employee_Number) as MX
                      FROM Staff_mx
                      GROUP BY Extract_Month) Sm ON S.Extract_Month = Sm.Extract_Month;

我在这里使用FULL OUTER JOIN只是，以防一个表在给定的提取月份中没有任何行。

【讨论】：