Question

任务：从MySQL标准数据库“世界”中选择人口超过一百万的所有城市，将国家数量限制为10个。如果我这样做：

SELECT country.Name, city.Name, city.Population
FROM country
INNER JOIN city
  ON country.Code = city.CountryCode
WHERE city.Population >= 1000000
LIMIT 10;

它产生了这个：

+-------------+--------------+------------+
| Name        | Name         | Population |
+-------------+--------------+------------+
| Afghanistan | Kabul        |    1780000 |
| Algeria     | Alger        |    2168000 |
| Angola      | Luanda       |    2022000 |
| Argentina   | Buenos Aires |    2982146 |
| Argentina   | La Matanza   |    1266461 |
| Argentina   | CÃ³rdoba     |    1157507 |
| Armenia     | Yerevan      |    1248700 |
| Australia   | Sydney       |    3276207 |
| Australia   | Melbourne    |    2865329 |
| Australia   | Brisbane     |    1291117 |
+-------------+--------------+------------+

例如，阿根廷重复了3次。我需要将国家/地区数量限制为10，而不是行数。由于我需要所有城市，因此GROUP BY不会这样做，我不需要将它们分组。

预期结果如下：

+--------------------+----------------------------+------------+
| Name               | Name                       | Population |
+--------------------+----------------------------+------------+
| Afghanistan        | Kabul                      |    1780000 |
| Algeria            | Alger                      |    2168000 |
| Angola             | Luanda                     |    2022000 |
| Argentina          | Buenos Aires               |    2982146 |
| Argentina          | La Matanza                 |    1266461 |
| Argentina          | CÃ³rdoba                   |    1157507 |
| Armenia            | Yerevan                    |    1248700 |
| Australia          | Sydney                     |    3276207 |
| Australia          | Melbourne                  |    2865329 |
| Australia          | Brisbane                   |    1291117 |
| Australia          | Perth                      |    1096829 |
| Azerbaijan         | Baku                       |    1787800 |
| Bangladesh         | Dhaka                      |    3612850 |
| Bangladesh         | Chittagong                 |    1392860 |
| Brazil             | SÃ£o Paulo                 |    9968485 |
| Brazil             | Rio de Janeiro             |    5598953 |
| Brazil             | Salvador                   |    2302832 |
| Brazil             | Belo Horizonte             |    2139125 |
| Brazil             | Fortaleza                  |    2097757 |
| Brazil             | BrasÃlia                   |    1969868 |
| Brazil             | Curitiba                   |    1584232 |
| Brazil             | Recife                     |    1378087 |
| Brazil             | Porto Alegre               |    1314032 |
| Brazil             | Manaus                     |    1255049 |
| Brazil             | BelÃ©m                     |    1186926 |
| Brazil             | Guarulhos                  |    1095874 |
| Brazil             | GoiÃ¢nia                   |    1056330 |
| United Kingdom     | London                     |    7285000 |
| United Kingdom     | Birmingham                 |    1013000 |
+--------------------+----------------------------+------------+

如您所见，无论有多少个城市，国家数量都是10个。

当然，我的数据库是不同的。为了简单，通用和可用性，我使用了“世界” 在真正的数据库中，我也对“country”进行了一些过滤。

Answer 1

我希望这会起作用

SELECT country.Name, city.Name, city.Population
FROM country
LEFT JOIN city
  ON country.Code = city.CountryCode
WHERE city.Population >= 1000000
AND city.id IN (SELECT id FROM city WHERE city.Population >= 1000000 AND country.Code = city.CountryCode LIMIT 10);

Answer 2

//So sorry try this query

SELECT country.Name, city.Name, city.Population
FROM country
LEFT JOIN city
  ON country.Code = city.CountryCode
WHERE city.Population >= 1000000 and country.Code IN(SELECT distinct CountryCode FROM city WHERE Population >= 1000000 LIMIT 10);

Answer 3

为子查询中的每个唯一国家/地区生成临时行级别ID，并使用where temp_row_num is <= 10等where条件从中选择行。

示例：

select
       /*uq_country_row_num, */ 
       country_name, city_name, population
from (
  select country_name, city_name, population
    , case when @pn=(@cn:=country_name) 
           then @r:=(@r + 1) 
           else @r:=1
      end as uq_country_row_num
    , @pn:=@cn as temp_cn
  from population, 
       (select @pn:='', @cn:='', @r:=0) country_rown_nums
  -- use order by country_name if the result set is un-ordered
  order by
     country_name
) all_results
where
  uq_country_row_num <= 10
;

如果您希望将变量/*uq_country_row_num, */作为输出的一部分，则可以取消注释。

示例 @ SQL Fiddle

修改：来自评论：

您的意思是每个国家/地区10行或最多10个国家/地区的所有记录？

确切地说，“来自最多10个国家/地区的所有记录”。

select country_id, country_name, city_name, population from ( select country_name, city_name, population , case when @pn=(@cn:=country_name) then @country_id:=@country_id else @country_id:=(@country_id + 1) end as country_id , @pn:=@cn as temp_cn from population, (select @pn:='', @cn:='', @row_num:=0, @country_id:=0) country_row_nums order by country_name ) all_results where country_id <= 10 ;

示例2 @ SQL Fiddle

Answer 4

尝试一次

SELECT city.Name, city.Population, newcountry.name
FROM city
INNER JOIN 
(select distinct Name,Code
 from country limit 10) newcountry
  ON newcountry.Code = city.CountryCode
WHERE city.Population >= 1000000;

Answer 5

没有一个答案奏效。所以我不得不使用以下请求：

SELECT co.Code, co.Name, ci.Name, ci.Population
  FROM city ci
    INNER JOIN (
      SELECT DISTINCT co.Code, co.Name
        FROM country co
          INNER JOIN city ci
          ON ci.countryCode = co.Code
        WHERE ci.Population > 1000000
        ORDER BY co.Name
        LIMIT 10
    ) co
    ON ci.countryCode = co.Code
  WHERE ci.Population > 1000000;

对我来说似乎相当笨拙而且效率不高（注意条件的重复），但由于缺乏更好的方法，这是唯一的解决方案。

感谢大家的回答。

MySQL SELECT。 LIMIT来自一列的值的数量

5 个答案: