hive 案例

创建一个alice 表：

create table alice(line string);

导入alice-in-wonderland.txt

统计每一行有多少个单词

     select split(line,' ') from alice where line!='' limit 3;

统计每一行有多少个单词

    select size(split(line,' ')) from alice;

    

但是其中包含很多空字符串,去掉空字符串
desc function array;
select * from alice where line!='aa';

select sum(size(split(line,' '))) from alice where line!='';

在hive中：函数分为:
udf:单行函数，一条记录一个返回结果
udaf：聚合函数

udtf：一条数据进去，多条记录出来,如explode

排序：

with av as (select explode(split(line,' ')) as one from alice 
where line!='')

select av.one,count(1) as one_num from av group by one order by one_num desc limit 50;