Oracle多行列值拼接并保留原行的SQL实现求助

可以使用分析函数版的LISTAGG直接实现需求，无需单独聚合后关联，在原表基础上即可生成每个PID对应的完整拼接文本：

SELECT 
    pid,
    LISTAGG("Desc", ' ') WITHIN GROUP (ORDER BY seq) OVER (PARTITION BY pid) AS Desc1,
    "Desc" AS Desc2
FROM A
ORDER BY pid, seq;

说明： #

• LISTAGG(...) OVER (PARTITION BY pid)：将每个PID分组内的所有Desc字段按seq顺序拼接，作为每条记录的Desc1
• 直接保留原表的Desc字段作为Desc2
• 最后按pid和seq排序，确保结果顺序与目标输出一致

执行该语句后，输出结果与目标格式完全匹配（注：原表中A的第三条Desc为day.，目标输出中误写为day!，执行结果将以原表数据为准）：

PID Desc1 Desc2
A Have a nice day. Have
A Have a nice day. a nice
A Have a nice day. day.
B Nice Work. Nice Work.
C Yes we can do this work! Yes
C Yes we can do this work! we can
C Yes we can do this work! do
C Yes we can do this work! this work!

内容的提问来源于stack exchange，提问作者Arpit Shukla