对聚合函数求和

Question

我试图通过使用嵌套的 select 语句对聚合函数中的总和进行求和，然后我在最后有一个分组依据，但我收到一个错误，关键字是预计不会。

代码如下：

SELECT SUM(TOTAL) as CURTOTAL, SUM(LYTOTAL) as CURLYTOTAL
from
(select 
pf.*, ps.*, st.*, RANK() OVER (ORDER BY JULsaless desc) as TOTRANK,
 JANSALESS + FEBSALESS + MARSALESS + APRSALESS + MAYSALESS + JUNSALESS + JULSALES as TOTAL,
JANLYSAL + FEBLYSAL + MARLYSAL + APRLYSAL + MAYLYSAL + JUNLYSAL + JULLYSAL as LYTOTAL` 
from 
payssfile pf left joinpayssspec ps on pf.str#` = ps.str# and pf.item# = ps.item#  join storefile 
st on
pf.str# = st.str# where(year = 2015 and totaltype = '' and pf.str# =38))` group by pf.str#;`

然后当我去运行时收到这条消息：

The keyword BY was not expected here. A syntax error was detected at keyword BY. The partial list of valid tokens is FOR USE SKIP WAIT WITH FETCH ORDER UNION EXCEPT OPTIMIZE.

Answer 1

无论您的 DBMS 是什么，您都缺少一个别名。您应该始终为您正在使用的派生 table 指定一个别名。

派生示例 table

select...
from (
-- this part of the query is a derived table:
select ... from ...
) as <ALIAS_HERE> -- must give derived table an alias

这就是您在 GROUP BY

附近出现错误的原因

在您的特定示例中，只需替换

)) group by pf.str#;

和

)) foo group by pf.str#; 
-- here you will name your derived table foo, to be able to call it's columns by foo.column_name

这是关于 SO 的类似问题的reference。

Answer 2

您似乎没有为子查询指定别名，pf 是子查询中 table 的名称，但不是子查询本身。因此，按功能分组被破坏。

对聚合函数求和

Summing a aggregate function

sql

db2

aggregate-functions