将许多独特的行合并为一个

Question

我正在努力创建一个选择查询，该查询将每一行合并/聚合为按列 postfix_id 分组的单行。

|   postfix_id  |   initial_datetime  |    final_datetime   |  from_address  |   to_address   | response |

|   655CB4B10   | 2017-06-30 00:17:13 |         null        |      null      |      null      |   null   |
|   655CB4B10   |         null        |         null        | bob@domain.tld |      null      |   null   |
|   655CB4B10   |         null        |         null        |      null      | sam@domain.tld |   null   |
|   655CB4B10   |         null        | 2017-06-30 00:17:14 |      null      |      null      |   sent   |
|   C32AE57F3   | 2017-06-10 07:14:36 |         null        |      null      |      null      |   null   |
|   C32AE57F3   |         null        |         null        | joe@domain.tld |      null      |   null   |
|   C32AE57F3   |         null        |         null        |      null      | tye@domain.tld |   null   |
|   C32AE57F3   |         null        | 2017-06-10 07:22:54 |      null      |      null      |  bounce  |

每个 postfix_id 值被列出 4 次，我需要将每个匹配项合并到一行中，从而删除剩余列中存在的空值。

例如所需的输出：

|   postfix_id  |   initial_datetime  |    final_datetime   |  from_address  |   to_address   | response |

|   655CB4B10   | 2017-06-30 00:17:13 | 2017-06-30 00:17:14 | bob@domain.tld | sam@domain.tld |   sent   |
|   C32AE57F3   | 2017-06-10 07:14:36 | 2017-06-10 07:22:54 | joe@domain.tld | tye@domain.tld |  bounce  |

该表有超过一百万个独特的 postfix_id，我浏览过其他类似的帖子，但是，它们有要比较的列和/或不相关的列。我无数次使用 MAX 和 GROUP_CONCAT 的尝试都失败了。

任何帮助/指导将不胜感激。我觉得好像很简单，导致 /facepalm

---

再次感谢，我最终得到了以下内容以满足我的最终需求。我希望这对以后的人有所帮助..

SELECT postfix_id,from_address,to_address,remote_response,insert_datetime,initial_datetime,final_datetime
FROM (
SELECT
postfix_id,
MAX(from_address)     AS from_address,
MAX(to_address)       AS to_address,
MAX(remote_response) AS remote_response,
MAX(insert_datetime)   AS insert_datetime,
MAX(initial_datetime) AS initial_datetime,
MAX(final_datetime)   AS final_datetime
FROM email_outbound_postfix
GROUP BY
postfix_id
) as t1
WHERE insert_datetime >= DATE_SUB(NOW(),INTERVAL 1 DAY)
AND CONCAT(from_address, ' ', to_address) LIKE '%sam%';

Answer 1

如果您向我们展示的数据准确无误，那么您只需在postfix_id 列上汇总您的表格，然后选择其他所有列的MAX()：

SELECT
    postfix_id,
    MAX(initial_datetime) AS initial_datetime,
    MAX(final_datetime)   AS final_datetime,
    MAX(from_address)     AS from_address,
    MAX(to_address)       AS to_address
FROM yourTable
GROUP BY
    postfix_id

这应该起作用的原因是MAX() 忽略 NULL 列中的值。所以假设每个postfix_id 组只有一个非NULL 值，那么MAX() 会选择它。

输出：

演示在这里：

Rextester

Answer 2

为此你可以简单地使用Mysql的MAX

 SELECT postfix_id , MAX(initial_datetime) AS initial_datetime ,MAX(final_datetime) AS final_datetime ,MAX(from_address) AS  from_address ,MAX(to_address) AS to_address,MAX(response) AS response from table group by postfix_id 

有关 MAX 的更多信息，请阅读https://dev.mysql.com/doc/refman/5.7/en/example-maximum-column.html