0

我创建了以下选择：

select e.first_name, e.last_name, 
   (select jobname_id from 
       (select j.id, j.jobname_id, first_value(j.jobname_id) 
        over (order by j.date_from desc) as current_job 
        from jobs j where j.emp_id=e.emp_id and j.date_from < sysdate) 
    where jobname_id != current_job and rownum=1) as previous_job 
from employees e

但我得到“ORA-00904：“E”。“EMP_ID”：无效标识符”

如何在子查询中使用对 E.EMP_ID 的引用？

4

2 回答 2

1

你不能，但你可以将连接移到外部子查询中。不过，不确定它是否适用于分析功能。

select e.first_name, e.last_name, 
   (select jobname_id from 
       (select j.emp_id, j.id, j.jobname_id, first_value(j.jobname_id) 
        over (order by j.date_from desc) as current_job 
        from jobs j where j.date_from < sysdate) j
    where j.emp_id=e.emp_id and jobname_id != current_job and rownum=1) as previous_job 
from employees e
于 2011-09-21T08:29:20.697 回答
1

我不知道这个解决方案是否足够快以满足您的需求，但它应该按计划工作。另外，我同意 Ollie 关于附加字段“date_to”的看法。这会让事情变得更容易。

select
  e.first_name   as first_name,
  e.last_name    as last_name,
  sub.jobname_id as jobname_id
from 
  employees e,
  (
    select
      j.emp_id     as emp_id,
      j.jobname_id as jobname_id,
      row_number()
        over (
          partition by j.emp_id
          order by     j.date_from desc
        ) as job_order
    from
      jobs j
    where j.date_from < sysdate
  ) sub
where sub.emp_id    = e.emp_id
  and sub.job_order = 2
于 2011-09-21T10:47:47.440 回答