Java 关于oracle sql group by'；不是按表达式分组'；

Oracle 10g，sql抛出异常：“不是按表达式分组”

然后，我更改了“分组依据”参数，添加了“h.hbsag”“h.sgpt”，如：

group by h.personal_info_id,h.hbsag,h.sgpt

---

但是结果不正确。

您应该添加不属于聚合函数的所有列。试试这个

select count(*) from 
( 
    select h.personal_info_id pid,h.hbsag hbsag,h.sgpt sgpt,MAX(h.date_reported) 
    from health_checkup_info h inner join personal_info p on h.personal_info_id = p.id 
    where h.deleted = 0 and h.date_reported is not null and h.hbsag in(1,2) and p.deleted = 0
    group by h.personal_info_id ,h.hbsag ,h.sgpt 
) t 

---

根据Oracle的行为，除聚合函数外的所有选定列都必须按分组

select count(*)
        from (
            select h.personal_info_id pid,h.hbsag hbsag,h.sgpt sgpt,MAX(h.date_reported)
            from health_checkup_info h
            inner join personal_info p on h.personal_info_id = p.id
            where
                h.deleted = 0
                and h.date_reported is not null
                and h.hbsag in(1,2)
                and p.deleted = 0

            group by h.personal_info_id,h.hbsag,h.sgpt
        ) t

更新： 由于您只是在使用

计数

，需要什么来获取其他列？？试试这个

select count(*)
        from (
            select h.personal_info_id pid,MAX(h.date_reported)
            from health_checkup_info h
            inner join personal_info p on h.personal_info_id = p.id
            where
                h.deleted = 0
                and h.date_reported is not null
                and h.hbsag in(1,2)
                and p.deleted = 0

            group by h.personal_info_id
        ) t;

---

谢谢大家，现在我已经解决了这个问题，问题是：

select count(*)
        from (
            select h.personal_info_id pid,h.hbsag,ROW_NUMBER() OVER (partition by h.personal_info_id order by h.date_reported desc) r
            from health_checkup_info h
            inner join personal_info p on h.personal_info_id = p.id
            where
                h.deleted = 0
                and h.date_reported is not null
                and h.hbsag in(1,2)
                and p.deleted = 0

        ) t where t.hbsag=2 and r=1

---

您不需要在GROUPBY子句中指定列别名。我想这是复制/粘贴的东西：）@Madhivanan，谢谢你，伙计。你明白了：）修好了@曼努拉贾达，谢谢你！但如果使用“h.personal\u info\u id、h.hbsag、h.sgpt分组”，则结果不正确。例如，使用“按h.personal\u info\u id分组”，结果是2640。但是，使用“按h.personal\u info\u id分组，h.hbsag，h.sgpt”，结果是2641。为什么？是的，让我更新查询，这只是实现目标的另一个逻辑。@manurajhada不，不，不。我需要h.hbsag，h.sgpt，然后再次使用它。“按h.personal_info_id）t分组，其中h.hbsag=1和h.sgpt>=20”

select count(*)
        from (
            select h.personal_info_id pid,h.hbsag,ROW_NUMBER() OVER (partition by h.personal_info_id order by h.date_reported desc) r
            from health_checkup_info h
            inner join personal_info p on h.personal_info_id = p.id
            where
                h.deleted = 0
                and h.date_reported is not null
                and h.hbsag in(1,2)
                and p.deleted = 0

        ) t where t.hbsag=2 and r=1