Question

我正在尝试列出全名和全职处理的交易总数员工，然后按交易总数的降序对结果进行排名。

这就是我所拥有的

select 
    concat( e.efirst, e.elast ) ename, 
    count(*) total_transactions
from 
    transactions t 
    join employees e on t.employeeid = e.employeeid
where 
    e.etype = 'Fulltime' 
group by
    t.employeeid 
order by
    total_transactions desc;

Answer 1

在通过外键对聚合进行计算时，您需要在外部查询中JOIN的主体关系（在这种情况下为employee），因为它是与聚合分开的关注点。

SELECT
    CONCAT( CONCAT( employee.efirst, ' ' ), employee.elast ) AS employee_name, 
    employee_id,
    total_transactions
FROM
    (
        SELECT
            employeeid AS employee_id,
            COUNT(*) AS total_transactions
        FROM
            transactions
        GROUP BY
            employeeid
    ) AS sq
    INNER JOIN employee ON sq.employee_id = employee.employeeid
WHERE
    employee.etype = 'Fulltime'
ORDER BY
    total_transactions desc;

Answer 2

添加concat(e.efirst, e.elast)进行分组：

select 
    concat(e.efirst, e.elast) ename, 
    count(*) total_transactions
from 
    transactions t 
    join employees e on t.employeeid = e.employeeid
where 
    e.etype = 'Fulltime' 
group by
    t.employeeid, concat(e.efirst, e.elast) 
order by
    total_transactions desc;

Answer 3

group by表达式应与select中未聚合的表达式匹配，因此：

select ( e.efirst || ' ' || e.elast ) as ename, 
       count(*) total_transactions
from transactions t join
     employees e
     on t.employeeid = e.employeeid
where e.etype = 'Fulltime' 
group by ( e.efirst || ' ' || e.elast ) 
order by total_transactions desc;

我还在名称之间添加了一个空格，并切换为使用串联运算符。

这假设员工姓名是唯一的。如果不是这种情况，那么您应该在select中加入员工ID，这样您就可以在同名员工中进行区分。

简单错误“ ORA-00979：不是GROUP BY表达式”

3 个答案: