SYNTAX_ERROR：'“ xyz”'必须是一个聚合表达式或出现在GROUP BY子句中

Posted 2021-03-27

我有两个表main_table和staging_table，main_table包含原始数据，而staging_table包含一些我必须与main_table数据一起添加的更新记录，为此，我使用的是唯一ID-PersonID和到达时间-date下面是我可以执行到SQL中的查询

SELECT PersonID, LastName, FirstName, Address, City, max(date) 
from 
(
select PersonID, LastName, FirstName, Address, City, date from main_table
UNION
select PersonID, LastName, FirstName, Address, City, date from staging_table
) as t
GROUP by t.PersonID;

但是在执行到AWS Athena时，出现以下错误，SYNTAX_ERROR：'“ LastName”'必须为聚合表达式或出现在GROUP BY子句中

答案

我怀疑其他列可能会有所不同，您实际上想要的是最近日期的完整记录。在这种情况下，请使用row_number()：

select p.*
from (select p.*,
             row_number() over (partition by personid order by date desc) as seqnum
      from ((select PersonID, LastName, FirstName, Address, City, date
             from main_table
            ) union all
            (select PersonID, LastName, FirstName, Address, City, date
             from staging_table
            )
           ) p
     ) p
where seqnum = 1;

此PersonId选择具有最新日期的一行。列来自最新行。

另一答案

您必须这样分组：

SELECT PersonID, LastName, FirstName, Address, City, max(date) 
from 
(
select PersonID, LastName, FirstName, Address, City, date from main_table
UNION
select PersonID, LastName, FirstName, Address, City, date from staging_table
) as t
GROUP by PersonID, LastName, FirstName, Address, City;

select子句中不属于聚合函数的每一列都必须位于order by子句中。

或者您可以这样做：

SELECT PersonID, max(LastName), max(FirstName), max(Address), max(City), max(date) 
from 
(
select PersonID, LastName, FirstName, Address, City, date from main_table
UNION
select PersonID, LastName, FirstName, Address, City, date from staging_table
) as t
GROUP by PersonID, LastName, FirstName, Address, City;

而且还有很多其他方式可以根据您需要的结果来编写代码。这个答案是对您的问题的答案：

SYNTAX_ERROR：'“ xyz”'必须是聚合表达式或出现在GROUP BY子句