错误：必须对属性进行分组或在聚合函数 NETEZZA SQL 中

Posted 2023-05-08

【中文标题】错误：必须对属性进行分组或在聚合函数 NETEZZA SQL 中

【英文标题】：Error: Attribute must be grouped or in aggregate function NETEZZA SQL

【发布时间】：2020-12-29 11:45:49

【问题描述】：

我正在编写 SQL 来创建如下的日历表并收到错误“属性 T1.CALENDAR_DATE 必须分组或在聚合函数中使用

我写的代码是

WITH CALENDAR AS
(
SELECT
EXTRACT(EPOCH FROM DATE(NOW()-1)-T1.CALENDAR_DATE)/7 AS ROLLING_WEEK,
MIN(T1.WEEK_OF_YEAR) AS WEEK,
MAX(T1.CALENDAR_DATE) AS LAST_DAY_WEEK,
MIN(T1.CALENDAR_DATE) AS FIRST_DAY_WEEK

FROM DIM_GINZ_DATE AS T1
WHERE T1.ROLLING_MONTH BETWEEN 0 AND 84
GROUP BY T1.ROLLING_WEEK
)

SELECT *
FROM CALENDAR

这里有什么问题？

谢谢:)

【问题讨论】：

尝试使用：group by EXTRACT(EPOCH FROM DATE(NOW()-1)-T1.CALENDAR_DATE)/7。

【参考方案1】：

您应该在GROUP by 中使用T1.CALENDAR_DATE 或在EXTRACT(EPOCH FROM DATE(NOW()-1)-T1.CALENDAR_DATE)/7 上使用一些聚合函数

WITH CALENDAR AS
(
SELECT
EXTRACT(EPOCH FROM DATE(NOW()-1)-T1.CALENDAR_DATE)/7 AS ROLLING_WEEK,
-- something like this need to be implemented if you dont want to group by T1.CALENDAR_DATE
-- EXTRACT(EPOCH FROM DATE(NOW()-1)-MIN(T1.CALENDAR_DATE))/7 AS ROLLING_WEEK 
MIN(T1.WEEK_OF_YEAR) AS WEEK,
MAX(T1.CALENDAR_DATE) AS LAST_DAY_WEEK,
MIN(T1.CALENDAR_DATE) AS FIRST_DAY_WEEK

FROM DIM_GINZ_DATE AS T1
WHERE T1.ROLLING_MONTH BETWEEN 0 AND 84
GROUP BY T1.ROLLING_WEEK
, T1.CALENDAR_DATE -- like this
-- or use EXTRACT(EPOCH FROM DATE(NOW()-1)-T1.CALENDAR_DATE)/7
)

SELECT *
FROM CALENDAR

如果select 子句中没有未聚合的列（如果使用），则它们不能在select 子句中使用。