Question

我无法弄清楚对此的查询，我发现的一切都是立即行不通的，因为＆＃34; group by＆＃34;方面。

ID   |    Username   | Age | arbitrary
1    |      bob      | 10  | 34
2    |      bob      | 10  | 54
3    |     alice     | 10  | 123

ID 1和2都会被删除，并且不会有任何改变。我设法创建了一个类似这样的问题的网络应用，并希望应用一个唯一的（用户名，年龄）来修补它，但由于重复，我无法做到。

Answer 1

DELETE FROM thetable dd
WHERE EXISTS (
    SELECT *
    FROM thetable xx
    WHERE xx.username = dd.username
    AND xx.age = dd.age
    AND xx.id <> dd.id   -- if you want to delete ALL duplicates
    -- AND xx.id < dd.id -- if you want to keep only ONE of the duplicates
    );

Answer 2

绕过＃34;＆＃34;方面你可以使用窗口函数：

with duplicates as (
   select id, 
          count(*) over (partition by username, age) as dup_count
   from the_table
)
select *
from duplicates
where dup_count > 1;

有关手册中窗口功能的更多信息：
http://www.postgresql.org/docs/current/static/tutorial-window.html

这可以与delete声明结合使用：

with duplicates as (
   select id, 
          count(*) over (partition by username, age) as dup_count
   from the_table
)
delete from the_table
where id in (select id 
             from duplicates 
             where dup_count > 1);

以上使用数据修改CTE。有关详细信息，请参阅手册：
http://www.postgresql.org/docs/current/static/queries-with.html#QUERIES-WITH-MODIFYING

SQLFiddle示例：http://sqlfiddle.com/#!15/8b0e3/1

Answer 3

DELETE FROM tablename 
WHERE username IN (SELECT username
              FROM (SELECT row_number() OVER (PARTITION BY username), username
                       FROM tablename) x 
             WHERE x.row_number > 1);


        ------RESULT------

 ID   |    Username   | Age | arbitrary
 -----+---------------+-----+----------
 3    |     alice     | 10  |   123

PostgreSQL查询删除具有两个重复列的所有结果

3 个答案: