hive 运行sql报错Expression Not In Group By Key
案例:
Select tmp.username,date,sum(tmp.su) over(partition by tmp.username order by tmp.date ) totle From ( Select username,sum(cost_money) su,date From table_test1 Group by date,username ) tmp Group by tmp.date,tmp.username Order by tmp.username;
mysql运行结果:
hive提示:Semantic Exception: Line 1:24 Expression not in GROUP BY key 'su' (state=42000,code=40000)
原因:1.Hive不允许直接访问非group by字段;
解决:
1.对于非group by字段,可以用Hive的collect_set函数收集这些字段,返回一个数组;使用数字下标,可以直接访问数组中的元素,比如:collect_set(username)[0];
2. first() 函数;注:first() 函数是分组第一个元素,与之对应的还有 last(),是分组最后一个元素。
也可以直接将字段放入group by中:
Select tmp.username,date,sum(tmp.su) over(partition by tmp.username order by tmp.date ) totle From ( Select username,sum(cost_money) su,date From table_test1 Group by date,username ) tmp Group by tmp.date,tmp.username,su Order by tmp.username;