Postgresql-在另一个select语句中使用变量

作为学习psql任务的一部分，我尝试将查询优化95%。我想得到的是所有拥有与第三大借书量的人相等或更高的客户

explain analyze
select clb.f_name, clb.l_name, noofbooks 
from 
      (select f_name, l_name, count(*) as noofbooks 
      from customer left join loaned_book on customer.customerid = loaned_book.customerid 
      group by f_name, l_name) as clb,
      (select f_name as fname, l_name as lname, count(*) as noofbooksf 
      from customer left join loaned_book on customer.customerid = loaned_book.customerid
      group by fname, lname order by noofbooksf  desc  LIMIT 1 OFFSET 2) as clb1
where noofbooks  >= noofbooksf 
order by noofbooks desc; 

这就是我到目前为止所做的，它被优化到80度，但是两个连接增加了成本。因此，我想做的只是做一次连接。然后使用结果得到它的第三个元素。我曾想过创建一个临时视图，但我相信会有更好的方法

---

谢谢。

如果你想要借书最多或借书第二的人，那么只需：

select c.f_name, c.l_name, count(lb.customerid) as noofbooks 
from customer c left join
     loaned_book lb
     on c.customerid = lb.customerid 
group by c.f_name, c.l_name
order by noofbooks desc
limit 3;

如果您关心领带：

select f_name, l_name, noofbooks
from (select c.f_name, c.l_name, count(lb.customerid) as noofbooks,
             rank() over (order by noofbooks desc) as seqnum
      from customer c left join
           loaned_book lb
           on c.customerid = lb.customerid 
      group by c.f_name, c.l_name
     ) clb
where seqnum <= 3;

---

等于或大于什么？例如，借书量排在第三位的人借了4本书。我们希望每个人都有4本或4本以上的借书。为什么要按f_名称、l_名称而不是客户ID进行分组？