Question

我有一个包含以下列的表：

id, name, domain, added, is_verified
1, "First Google", "google.com", DATE(), 1
2, "Second Google", "google.com", DATE(), 1
3, "Third Google", "google.com", DATE(), 1
4, "First disney", "disney.com", DATE(), 1
5, "Second disney", "disney.com", DATE(), 1
6, "Third disney", "disney.com", DATE(), 0
7, "First example", "example.com", DATE(), 0
8, "Second example", "example.com", DATE(), 0

以下要求：

SELECT domain FROM mytable WHERE domain NOT IN 
    (SELECT domain FROM mytable WHERE is_verified = 1 GROUP BY domain)
GROUP BY domain ORDER BY added DESC;

此请求背后的主要思想是将所有没有domain的{{1}}设置为true。

在上面的例子中，这只会返回“example.com”一次。

请求运行良好，但需要时间来执行（我有数千个条目）。还有其他方法可以使这个请求更快更有效吗？

Answer 1

您可以LEFT JOIN使用NULL支票：

SELECT T1.Domain
FROM mytable T1
LEFT JOIN mytable T2 ON T2.domain = T1.domain AND T2.is_verified = 1
WHERE T2.ID IS NULL

使用给定数据执行示例：

DECLARE @TESTDOMAIN TABLE (id int, name varchar(100), domain varchar (100),  added datetime, is_verified bit)

insert into @testdomain (id, name, domain, added, is_verified)
SELECT 1, 'First Google', 'google.com', GETDATE(), 1 UNION
SELECT 2, 'Second Google', 'google.com', GETDATE(), 1 UNION
SELECT 3, 'Third Google', 'google.com', GETDATE(), 1 UNION
SELECT 4, 'First disney', 'disney.com', GETDATE(), 1 UNION
SELECT 5, 'Second disney', 'disney.com', GETDATE(), 1 UNION
SELECT 6, 'Third disney', 'disney.com', GETDATE(), 0 UNION
SELECT 7, 'First example', 'example.com', GETDATE(), 0 UNION
SELECT 8, 'Second example', 'example.com', GETDATE(), 0

SELECT T1.Domain
FROM @TESTDOMAIN T1
LEFT JOIN @TESTDOMAIN T2 ON T2.domain = T1.domain AND T2.is_verified = 1
WHERE T2.ID IS NULL

Answer 2

SELECT domain 
FROM mytable 
group by domain 
having max(is_verified) = 0
ORDER BY max(added) DESC

我添加了order by子句。您必须决定要为每个域采用哪个added记录。我选择了域的最大附加值。

Answer 3

为什么必须使用子选择？不会产生相同的结果吗？

package ls2.repositories;

import org.springframework.data.neo4j.repository.GraphRepositoryImpl;
import org.springframework.data.neo4j.template.Neo4jOperations;

import ls2.models.User;

public class UserOpImpl extends GraphRepositoryImpl<User> implements UserOp {

    public UserOpImpl(Neo4jOperations neo4jOperations) {
        super(User.class, neo4jOperations);
    }

}

有没有更快的方法来执行以下SQL请求？

3 个答案: