在这种情况下如何删除重复的行？ [复制]

Question

|  id  |        name        |              date           |     points   |
|  10  |        Paul        |     2013-04-29 10:15:03     |       2      |
|  11  |       Joseph       |     2013-04-29 10:50:17     |       0      |
|  12  |       Joseph       |     2013-04-29 11:23:18     |       10     |
|  13  |        Bill        |     2013-04-29 11:27:10     |       8      |
|  14  |        Paul        |     2013-04-29 11:41:38     |       5      |
|  15  |       Joseph       |     2013-04-29 11:43:15     |       0      |
|  16  |       Joseph       |     2013-04-29 11:47:30     |       0      |
|  17  |       Joseph       |     2013-04-29 12:51:38     |       0      |
|  18  |       Joseph       |     2013-04-29 12:53:58     |       10     |
|  19  |        Bill        |     2013-04-29 13:17:10     |       8      |
|  20  |       Joseph       |     2013-04-29 13:21:38     |       7      |

只有寄存器 16 和 17 必须删除。

我需要的是，每次有来自同一用户的 0 序列时，都删除相同的序列，除了第一个 0，在本例中为 id 号 15。

Answer 1

假设date字段总是递增的，你可以按照下一个过程进行

1. 跟踪重复记录和这些记录的最短日期
2. 删除日期值大于最小日期的所有记录。

代码示例：

第 1 步：

select name, points, count(id) as rowCount, min(id) as minId, min(`date`) as minDate
from yourTable
where points = 0
group by name
having count(id)>1

第 2 步：

delete from yourTable
where id in (
    select id
    from yourTable
    inner join (
            select name, points, min(id) as minId, count(id) as rowCount, min(`date`) as minDate
            from yourTable
            where points = 0
            group by name
            having count(id) > 1
        ) as a on yourTable.name = a.name and yourTable.id > a.minId
    )
and points = 0;

希望对你有帮助

---

我认为使用临时表来获取要删除的 id 可能很有用：

-- Step 1: Create a temporary table with the names of the people you want to remove
drop table if exists temp_dup_names;
create temporary table temp_dup_names
    select name, points, min(id) as minId, count(id) as rowCount, min(`date`) as minDate
    from yourTable
    where points = 0
    group by name
    having count(id) > 1;
alter table temp_dup_names
    add index idx_name(name),
    add unique index idx_id(minId);

-- Step 2: Create a temporary table with the ids you want to delete
drop table if exists temp_ids_to_delete;
create temporary table temp_ids_to_delete
    select distinct a.id
    from yourTable as a
    inner join temp_dup_names as b on a.name=b.name and a.id > b.minId
    where points = 0;
alter table temp_ids_to_delete
    add unique index idx_id(id);

-- Step 3: Delete the rows
delete from yourTable
where id in (select id from temp_ids_to_delete);
-- If MySQL is configured in 'safe mode', you may need to add this 
-- to the where condition:
-- and id > 0;