Question

如果可能的话，如果可能的话，我可以通过表格中给定用户和每个其他用户之间的匹配数来“得分”结果。

基本上，我有一个'1'的'userid'，我需要根据用户兴趣和订单检查所有这些用户的兴趣，以及用户之间匹配的数量。

假设userid'1'与userid'4'有4个共同兴趣，但只有3个与userid'2'共同的兴趣，它将首先返回4，然后是2，然后整体兴趣较少的其他所有人匹配。< / p>

我已经尝试了大约10个不同的查询无效，甚至没有一个接近，这就是为什么我没有在这里发布任何尝试。

CREATE TABLE `interests` (
  `id` int(12) NOT NULL AUTO_INCREMENT,
  `userid` int(12) NOT NULL DEFAULT '0',
  `interest` varchar(100) NOT NULL DEFAULT '',
  PRIMARY KEY (`id`)
) ENGINE=InnoDB;

INSERT INTO `interests` set `userid` = '1', `interest` = 'term1';
INSERT INTO `interests` set `userid` = '1', `interest` = 'term2';
INSERT INTO `interests` set `userid` = '1', `interest` = 'term3';
INSERT INTO `interests` set `userid` = '1', `interest` = 'term4';
INSERT INTO `interests` set `userid` = '2', `interest` = 'term1';
INSERT INTO `interests` set `userid` = '2', `interest` = 'term2';
INSERT INTO `interests` set `userid` = '2', `interest` = 'term8';
INSERT INTO `interests` set `userid` = '2', `interest` = 'term4';
INSERT INTO `interests` set `userid` = '3', `interest` = 'term9';
INSERT INTO `interests` set `userid` = '3', `interest` = 'term2';
INSERT INTO `interests` set `userid` = '3', `interest` = 'term3';
INSERT INTO `interests` set `userid` = '3', `interest` = 'term7';
INSERT INTO `interests` set `userid` = '4', `interest` = 'term1';
INSERT INTO `interests` set `userid` = '4', `interest` = 'term2';
INSERT INTO `interests` set `userid` = '4', `interest` = 'term3';
INSERT INTO `interests` set `userid` = '4', `interest` = 'term4';

正确的结果集应如下所示：

-------
userid
-------
4
2
3

非常感谢大家的帮助。

Answer 1

您可以执行SELF JOIN：

SELECT i2.userid
FROM interests i1
INNER JOIN interests i2
  ON i2.userid <> i1.userid
  AND i2.interest = i1.interest
WHERE i1.userid = 1
GROUP BY i2.userid
ORDER BY COUNT(*) DESC;

ONLINE DEMO

Answer 2

最简单的方法是获取所有用户和共享和计数的列表，然后从该子查询中选择和排序所需的用户。它使逻辑清晰：

SELECT userid, otherid, sameCount
FROM (
  SELECT base.userid, other.userid as otherid, count(*) sameCount
  FROM interests base
  JOIN interests other ON base.interest = other.interest and base.userid != other.userid
  GROUP BY base.userid, other.userid
) sub
WHERE userid = 1
ORDER BY sameCount DESC

Answer 3

此查询应该这样做

select i.userid, count(i2.userid) countInterests
from interests i
inner join interests i2 on i.interest = i2.interest and i.Id <> i2.id
group by i.userid, i2.userid
order by userid, count(i2.userid) desc

如何获取按ID加入的多个表的匹配计数？

3 个答案: