我有一个查询应该读取leave_requests_2014
表中的记录,并为每位员工的所有记录执行sum()
个已接受的请假请求!
SELECT lr.emp_id employee,
ifnull(MAX(CASE WHEN lr.leave_type =2 THEN (sum((TO_DAYS(lr.end_date) - TO_DAYS(lr.start_date)+1))) END ) ,0) annual_leave,
ifnull(MAX(CASE WHEN lr.leave_type =3 THEN (sum((TO_DAYS(lr.end_date) - TO_DAYS(lr.start_date)+1))) END ),0) sick_leave
FROM leave_requests_2014 lr
WHERE lr.status =5
GROUP BY lr.emp_id
返回以下错误的查询:
#1111 - Invalid use of group function
我真的不知道应该分组什么来获得所需的输出!
请查看sqlfiddle这是我实际需要的结果
employee annual_leave sick_leave
5 7 2
6 4 1
答案 0 :(得分:1)
您的查询的更简单(和更正确)版本:
SELECT lr.emp_id AS employee,
SUM(IF(lr.leave_type = 2, DATEDIFF(lr.end_date, lr.start_date) + 1, 0)) AS annual_leave,
SUM(IF(lr.leave_type = 3, DATEDIFF(lr.end_date, lr.start_date) + 1, 0)) AS sick_leave,
FROM leave_requests_2014 lr
WHERE lr.status = 5
GROUP BY lr.emp_id
另一个版本,甚至更简单,可以在不同的行上生成年假和病假天数(如果需要):
SELECT lr.emp_id AS employee, lr.leave_type,
SUM(DATEDIFF(lr.end_date, lr.start_date) + 1) AS leave_days,
FROM leave_requests_2014 lr
WHERE lr.status = 5 AND lr.leave_type IN (2, 3)
GROUP BY lr.emp_id, lr.leave_type
这个需要对客户端代码进行一些处理:检查leave_type
的值,以了解leave_days
中的值是“annual_leave”还是“sick_leave”。如果查询未返回,则使用0
表示其他类型休假的天数
答案 1 :(得分:1)
"分组"诸如MAX()和/或SUM()之类的函数不能在同一语句中递归执行。解决方案是在子查询中执行第一个聚合。例如:
SELECT employee,
ifnull(MAX(CASE WHEN leave_type =2 THEN annual_leave END),0) annual_leave,
ifnull(MAX(CASE WHEN leave_type =3 THEN sick_leave END),0) sick_leave
FROM
(
SELECT lr.emp_id employee, lr.leave_type
(CASE WHEN lr.leave_type =2 THEN (sum((TO_DAYS(lr.end_date) - TO_DAYS(lr.start_date)+1))) END ) annual_leave,
(CASE WHEN lr.leave_type =3 THEN (sum((TO_DAYS(lr.end_date) - TO_DAYS(lr.start_date)+1))) END ) sick_leave
FROM leave_requests_2014 lr
WHERE lr.status =5
GROUP BY lr.emp_id, lr.leave_type
) firstAggregation
group by employee