Question

我正在处理需要检索的报告

所有员工
本月新增
今年新增
本月终止
今年终止

所有这些都被部门分开（新营销，新销售等）。

总共会有大约23列..

我通过使用更新填充每个字段来使用临时表。这是我正在做的简化示例。 Employee表具有所需的大部分值（hiredate，termdate等）。实际上每个更新都有一个连接，还有一些更多的条件。

update #tmptbl
set MtdHiresSales = select count(empid) from emp e where e.dept = 012 
                           and hiredate between start_of_month() and getdate()
                           -- more predicates
set MtdHiresMkting = ...repeat... with e.dept=013

我确信有更好的方法，因为有很多代码重复。在这种情况下临时表是否合适？我不确定没有一个怎么办。有什么建议吗？

Answer 1

您可以使用CASE。例如：

SELECT 
SUM(CASE WHEN e.dept=012 
    AND Hiredate between Start_of_Month() and Getdate() 
    AND... 
    THEN 1 ELSE 0 END) AS 'MtdHiresSales',
...
FROM ...
WHERE ...

Answer 2

您可以使用公用表表达式来封装员工计数代码...

WITH
  employee_counts
AS
(
   SELECT
     dept         AS "dept",
     x            AS "x",
     COUNT(empid) AS "employees"
   FROM
     emp
   WHERE
     hiredate BETWEEN start_of_month() AND getdate()
     -- etc, etc
   GROUP BY
     dept,
     x
)
UPDATE
  myTable
SET
  MtdHiresSales  = (SELECT employees FROM employee_counts WHERE dept = 012 AND x = myTable.y),
  MtdHiresMkting = (SELECT employees FROM employee_counts WHERE dept = 013 AND x = myTable.y)

Answer 3

也许我很困惑或没有看到什么，但为什么不只是使用聚合？

SELECT count(*) as employeesPerDept, deptId
from mudkips
group by deptId

给（我的样本数据）

employeesPerDept    deptId
6                   100
4                   200

和

select deptId, count(*) as empsHired, month(hiredate) as monthHired, 
year(hireDate) as yearHired
from mudkips
group by deptId, month(hireDate), year(hireDate)

然后是

deptId  empsHired   monthHired  yearHired
100         4           1           2010
100         1           1           2011
100         1           2           2010
200         2           1           2010
200         1           2           2010
200         1           3           2010

检索不同类型的月度数据

3 个答案: