2

我的数据库类有一个我无法弄清楚的问题。

我有一张名为counties的表，如下所示：

state_code | name | population
           |      |
           |      |

（每个 state_code 中有多个县）

还有一个叫做states的表，它只是

state_code | name 
           |
           |

无论如何，我需要一个返回方案（state_name，county_name，county_population）的查询，它按州名的顺序列出每个州中人口最多的五个县名（按人口降序排列）以及这些县的人口。

我有一个生成正确方案的查询，但它显示的结果太多，而不仅仅是前 5 个：

SELECT state.name AS state_name, county.name AS county_name, county.population 
FROM state JOIN county ON state.code = county.state_code 
GROUP BY state.name, county.name, county.population
ORDER BY state.name, county.population DESC

我尝试了涉及排名的更复杂的解决方案，但是我们学校只有 PostgreSQL 版本 8.3，它没有 PARTITION OVER 或 RANK()，这让事情变得非常困难。

谢谢

4

2 回答 2

2 
select state_name, county_name, population
from (
    select
        s.name state_name,
        c.name county_name,
        c.population,
        row_number() over(partition by s.state_code order by population desc) rn
    from
        states s
        inner join
        counties c on s.state_code = c.state_code
) s
where rn <= 5
order by state_name, population desc

row_number 窗口函数对每个状态内的行进行编号。在外部查询中，我将行号限制为 5 或更少。

于 2012-10-27T19:18:55.240 回答
0

这个解决方案非常缓慢和痛苦，但我相信它应该给你你想要的。我很想知道是否有人能想出一个更优雅的解决方案。

 SELECT state.name AS state_name, county.name AS county_name, county.population
 FROM state JOIN county ON state.code = county.state_code
 WHERE  county.name IN (SELECT county.name, COUNT(*) FROM county
          INNER JOIN county AS second_county ON (county.name = second_county.name) AND 
          (county.state_code = second_county.state_code) AND (county.population < state.population) GROUP BY county.name HAVING COUNT(*) <=5)
GROUP BY state.name, county.name, county.population
ORDER BY state.name, county.population DESC
于 2012-10-27T19:17:48.033 回答