分区以计算 "collapsed" 数值变量的百分比

Question

我有一个 table 年龄的人物数据。我想使用 SQL 计算我的 table 中三个或更多折叠年龄类别（例如 18 - 29、30 - 39、40 - 49 等）中的人数百分比。

基本上，我希望我的 table 看起来像这样：

Count | AgeCat | percent
400 | 50 and older | .40
300 | 35 to 49 | .30
300 | Under 35 | .30

这是我的查询：

select count(*), 
case when age >= 50 then '50 and older' when age < 50 and age >= 35 then '35 to 49' else 'Under 35' end as AgeCat,
count(*)/sum(count(*)) over (partition by '') as percent 
from mydata group by 2;

上面的 window 函数仅在我将我的年龄崩溃限制为两类崩溃（50 岁以上/50 岁以下）时才有效。

Answer 1

这是您的查询：

select count(*), 
       (case when age >= 50 then '50 and older'
             when age < 50 and age >= 35 then '35 to 49' 
             else 'Under 35'
        end) as AgeCat,
       count(*)/sum(count(*)) over (partition by '') as percent 
from mydata
group by 2;

这应该有效。但是，我会写成：

select count(*), 
       (case when age >= 50 then '50 and older'
             when age >= 35 then '35 to 49' 
             else 'Under 35'
        end) as AgeCat,
       ( count(*) / sum(count(*)) over () ) as percent 
from mydata
group by AgeCat
order by min(age);

分区以计算 "collapsed" 数值变量的百分比

Partition to calculate a percentage across a "collapsed" numeric variable

sql

vertica