Sql 按Apache配置单元分组时出错
我有一个表,其中包含:Sql 按Apache配置单元分组时出错,sql,hadoop,group-by,hive,Sql,Hadoop,Group By,Hive,我有一个表,其中包含: userid INT movieid INT rating FLOAT timestmp STRING select movieid, ROUND(AVG(rating),1) as Rating, COUNT(userid) as rtn_crt, ROUND(((Rating*rtn_cnt)+(100*3.5))/(rtn_cnt+100),1) as w_rating from ratings GROUP BY movieid LIMIT 50; 错误消息
userid INT
movieid INT
rating FLOAT
timestmp STRING
select movieid, ROUND(AVG(rating),1) as Rating, COUNT(userid) as rtn_crt, ROUND(((Rating*rtn_cnt)+(100*3.5))/(rtn_cnt+100),1) as w_rating
from ratings
GROUP BY movieid
LIMIT 50;
错误消息:
org.apache.hive.service.cli.HiveSQLException
:编译语句时出错:失败:SemanticException
[错误10025]:行2:6表达式不在GROUP BY keyrtn\U cnt
我尝试使用函数强制转换,但仍然不起作用,并收到相同的错误
select movieid, CAST(AVG(rating) AS FLOAT) as Rating, COUNT(userid) as rtn_crt,
CAST((Rating*rtn_cnt) AS FLOAT) + CAST((100*$AVG_MEAN) AS FLOAT)
/CAST((rtn_cnt+100) AS FLOAT) as w_rating
from ratings
GROUP BY movieid
LIMIT 50;
我建议您使用如下子查询重写SQL查询:
SELECT t.*, ROUND(((t.Rating*t.rtn_cnt)+(100*3.5))/(t.rtn_cnt+100),1) as w_rating
FROM (
SELECT movieid, ROUND(AVG(rating), 1) as Rating, COUNT(userid) as rtn_crt
FROM ratings
GROUP BY movieid
LIMIT 50
) t;