SQL DB2：在列值更改之前需要最新记录

需要选择属于在列ANDRI（value=p）中具有唯一值的客户机的所有最近记录，直到列ANDRI中的值发生变化（value=Q）。记录需要根据表B中的客户机名称ID（唯一ID）进行分组。两个表可以通过列ID（主键）连接

下面是我的DB2表A：

+---------+---------------+------+-----+---------+-------+
| ID      | Client Type   | ANDRI|  Date (Timestamp)     |
+---------+---------------+------+-----+---------+-------+
| 878     |  Personal     | P    | 2020-09-22 16:47:08   |
| 576     |  Personal     | P    | 2020-09-22 10:47:08   |
| 745     |  Personal     | P    | 2019-05-21 14:47:08   |
| 142     |  Business     | Q    | 2019-09-20 16:11:08   |
| 711     |  Personal     | Q    | 2018-02-12 15:27:08   |
| 441     |  Personal     | P    | 2018-01-29 10:57:08   |
| 371     |  Personal     | P    | 2017-05-20 11:17:08   |
| 115     |  Personal     | P    | 2016-10-12 14:47:08   |
+---------+---------------+------+-----+---------+-------+

这是我的DB2表B：

+---------+---------------+------+-----+---------+-------+
| ID      | Client Name ID| ODER|  Date (Timestamp)     |
+---------+---------------+------+-----+---------+-------+
| 878     |     Alice     | A    | 2020-09-22 16:47:08   |
| 576     |     Alice     | A    | 2020-09-22 10:47:08   |
| 745     |     Alice     | A    | 2019-05-21 14:47:08   |
| 142     |     Sandra    | B    | 2019-09-20 16:11:08   |
| 711     |     Alice     | B    | 2018-02-12 15:27:08   |
| 441     |     Alice     | A    | 2018-01-29 10:57:08   |
| 371     |     Sandra    | A    | 2017-05-20 11:17:08   |
| 115     |     Sandra    | A    | 2016-10-12 14:47:08   |
+---------+---------------+------+-----+---------+-------+

我预计产出如下：

| ID      | Client Name ID| Date (Timestamp)     |
+---------+---------------+------+-----+---------+
| 878     |  Alice        | 2020-09-22 16:47:08  |
| 576     |  Alice        | 2020-09-22 10:47:08  |
| 745     |  Alice        | 2019-05-21 14:47:08  |

如果可能的话，如果我们可以在SQL中使用ROW_NUMBER（）函数，这将非常有用

我一直在使用的SQL（但需要一些调整）：

---

您好@sam243475您有create table语句吗？您是否有insert into table语句？你自己尝试过任何代码吗？请发布你所有的代码。您需要自己编写SQL，我们对其进行了改进。添加了SQL代码如果客户端名称ID为1列，则需要引用名称，如果是3列，则需要在它们之间添加“，”。

SELECT T.*
FROM TABLEB T
JOIN 
(
    SELECT Client Name ID 
    FROM ( 
        SELECT Date, Client Name ID
             , ROWNUMBER() OVER (PARTITION BY Client Name ID 
                                 ORDER BY Date DESC) RN
        FROM TABLEB
    )
    WHERE RN IN (1, 2)
    GROUP BY Client Name ID
    HAVING MIN(ODER) = MAX(ODER) AND COUNT(1) = 2 AND MIN(ODER) = 'A'
) G 
    ON G.client name ID = T.client name ID
    AND T.ID IN (SELECT ID FROM TABLEA WHERE ANDRI = 'P')
ORDER BY T.Client Name ID;

select b.*
from a join
     b
     on a.id = b.id
where a.date > (select max(a2.date)
                from a a2
                where a2.andri <> 'P'
               );

select b.*
from b
where b.date > (select max(a2.date)
                from a a2
                where a2.andri <> 'P'
               );