Question

select season,violation_code, cnt, 
       RANK() over (Partition BY season order by cnt desc) AS rank 
  from
    (  select season,violation_code, 
              count(*) as cnt 
         from  ParkingViolations_seondary 
     group by season,violation_code
    ) tmp
where rank <= 3

我是Hive的新手。有人可以帮我以上查询有什么问题吗？它将引发以下错误：

编译语句时出错：

失败：SemanticException [错误10004]：第4：6行无效的表别名或列引用“等级” ：（可能的列名称为：季节，违规代码，cnt）

任何快速帮助将不胜感激。

Answer 1

使用子查询能够在rank子句中访问where：

select season, violation_code, cnt, rnk
from
( select season,violation_code, cnt, 
         RANK() over (Partition BY season order by cnt desc) AS rnk
    from
        (  select season,violation_code, 
                count(*) as cnt 
             from  ParkingViolations_seondary 
         group by season,violation_code
        ) tmp
)s
where rnk <= 3

Answer 2

是的，我还可以通过以下方法使它工作：

选择*从（ SELECT season，violation_code，cnt，RANK（）over（按季节划分ORDER BY cnt DESC）AS频率从（SELECT season，violation_code，COUNT（*）as cnt from ParkingViolations_seondary 其中（违反代码<> 0）和（街道代码1 <> 0或街道代码2 <> 0或街道代码3 <> 0）按季节分组，违反代码）TMP ）TMP1 频率<= 3;

Hive Rank查询问题

2 个答案: