sql - 使用子查询进行比较

Question

这是我试图回答的问题：

我所拥有的是：

SELECT DISTINCT a1.acnum
FROM academic a1, academic a2
WHERE a1.deptnum = a2.deptnum
AND a1.acnum <> a2.acnum
AND a1.acnum IN (Select acnum
from interest
group by acnum
having count(acnum) >1);

这是错误的，因为我所做的是如果 acnum（学号）218 与 acnum 217 在同一部门工作，并且与 acnum 199（差异部门）具有相同的兴趣，那么我将 acnum 218 添加到列表中。

但是，如果他们都有相同数量的领域兴趣，我应该只添加 acnum 218 和 217。

兴趣表有 fieldnum , acnum

学术表有 acnum 、 deptnum 、 name

部门表有deptnum，deptName

 FIELDNUM           ACNUM DESCRIP                                                                        
------------------ -------------------- 
292                 100  Multiprocessor and Special purpose computer design                               
293                 100  General (HW)                                                                     
293                 197  Computer architecture  



                                                      

输出应该只列出所有学者的号码.. 但要清楚一点：

Acnum Deptnum Interest
1        1       g&f
2        1       g&f
3        2        f
4        3        l
5        4       r&l
6        4       r&l

输出应该是：1 2 5 6

Answer 1

未经测试，但应该是好的

SELECT DISTINCT a1.acnum
FROM academic a1
INNER JOIN academic a2 ON a1.deptnum = a2.deptnum
                          AND 
                          a1.acnum <> a2.acnum
INNER JOIN interest i1 ON a1.acnum=i1.acnum
GROUP BY a1.acnum
HAVING COUNT(i1.acnum)=(SELECT COUNT(*)
                        FROM interest i2
                        WHERE i1.acnum=i2.acnum)

Answer 2

使用公用表表达式（子查询）来获取学者、他们的部门和他们的兴趣计数。然后从中查询两次以获得所需的输出。

with cte as ( select a.acnum
                     , a.deptnum
                     , count(i.acnum) as int_cnt
              from academic a
                   , interest i
              where i.acnum = a.acnum
              group by  a.acnum
                     , a.deptnum
            )
select ct1.acnum
       , cte1.deptnum
       , cte1.in_cnt
from cte cte1
     , cte cte2
where cte2.deptnum = cte1.deptnum 
and cte2.int_cnt = cte1.int_cnt  
and cte2.acnum != cte1.acnum 
order by cte1.deptnum
         , cte1.acnum

警告- 未经实际测试，因此虽然逻辑合理，但语法可能有误；）

Answer 3

也许JOIN在这里使用可以给你更好的结果：

SELECT DISTINCT a1.acnum
FROM academic a1
JOIN academic a2
ON a1.deptnum = a2.deptnum
AND a1.acnum <> a2.acnum
AND a1.acnum IN (Select acnum
from interest
group by acnum
having count(acnum) >1);

Answer 4

据我了解，它应该工作

SELECT 
      listagg(a1.acnum,',') within group( order by a1.acnum) , a1.deptnum,a2.cnt
FROM 
     academic a1,
     (Select 
            acnum,count(*) as cnt
      from interest
      group by acnum
     ) a2
where 
     a1.acnum=a2.acnum
group by 
     a1.deptnum,a2.cnt
having count(*)>1;