Question

我正在尝试采用该组中最常见的值进行汇总，如下所示：

with t1 as (
    select
        id
        , colA
        , colB
    from some_Table
)
select 
    id
    , count(*) as total
    , max(colA) as maxColA
    , most_common(colB) -- this is what I'm trying to achieve
from t1
group by id

这是我试图做的：

with t1 as (
    select
        id
        , colA
        , colB
    from some_Table
)
select 
    id
    , count(*) as total
    , max(colA) as maxColA
    , (select colB, count(colB) as counts from t1 group by colB order by counts desc limit 1) as most_freq_colB_per_id
from t1
group by id

但是，它告诉我AnalysisException: Subqueries are not supported in the select list。我还能怎么做？

Answer 1

据我所知，Impala没有内置的聚合函数来计算模式（您要计算的统计名称）。

您可以使用两个级别的聚合。您的CTE没有执行任何操作，因此您可以执行以下操作：

select id, sum(total) as total, max(maxColA) as maxColA,
       max(case when seqnum = 1 then colB end) as mode
from (select id, colB, count(*) as total, max(colA) as maxColA,
             row_number() over (partition by id order by count(*) desc) as seqnum
      from sometable
      group by id, colb
     ) t
group by id;

（Impala）在字段中选择最常见的值会导致“选择列表中不支持子查询”

1 个答案: