Hive SQL整理

collect_set

• 作用：把一个字段中的值组合到一个集合中，搭配where（筛选）、group by使用；
• 用途：可将非group by字段进行转换，使其可以查询；
• 示例：

-- id：用户；time：登陆日期
select id,collect_set(time) as t from t_action_login where time<='20150906' group by id

该查询结果如下：

  123@163.com                                                                                      | ["20150620","20150619"]                                                                                    |
| abc@163.com                                                                                      | ["20150816"]                                                                                               |
| cde@qq.com                                                                                       | ["20150606","20150608","20150607","20150609","20150613","20150610","20150616","20150615"]                  |
| 789@sohu.com                                                                                     | ["20150827","20150623","20150627","20150820","20150823","20150612","20150717"]                             |
| 987@163.com                                                                                      | ["20150701","20150829","20150626","20150625","20150726","20150722","20150629","20150824","20150716","20150 |
| ddsf@163.com                                                                                     | ["20150804","20150803","20150801","20150809","20150807","20150806","20150905","20150904","20150730","20150 |
| 182@163.com                                                                                      |["20150803","20150801","20150809","20150808","20150805","20150806","20150906","20150904","20150730","20150 |
| 22225@163.com                                                                                    | ["20150604","20150609","20150622","20150827","20150625","20150620","20150613","20150610","20150614","20150 |
| 18697@qq.com                                                                                     | ["20150902"]                                                                                               |
| 1905@qq.com                                                                                      | ["20150709"]                                       

该结果显示每个用户在2015年9月6日之前的全部登陆时间

• 场景应用：在hive中求出一个数据表中在2015年9月6日首次登陆的人数
  对应SQL语句：

select count(a.id) from (select id,collect_set(time) as t from t_action_login where time<='20150906' group by id) as a where size(a.t)=1 and a.t[0]='20150906'

注意：size(a.t)=1 and a.t[0]='20150906'

参考：https://blog.csdn.net/liyantianmin/article/details/48262109

group by

• hive不允许直接查询访问非group by的字段，错误示例：

select name,sex from people group by sex;
该语句会报错，因为name是非group by字段，解决方法如下：

1. 只查询group by字段：
   select sex from people group by sex;
2. 对于非group by字段，只通过聚合函数查询：
   select sex,count(name) from people group by sex;
3. 对于非group by字段，通过collect_set等函数转换后查询：
   select sex,collect_set(name)[0] as name from people group by sex;
   注：可以用原字段名称；

grouping sets

• 用途：可以按照多个标准进行分组，增强group by的功能；
• 示例：查询每个部门每类工作平均工资、每个部门的平均工资及整个企业中每个雇员的平均工资

select department_id,job_id,avg(salary)
from hr.employees
group by grouping sets ((department_id,job_id),department_id,()); 
-- ()表示对整体进行一次聚类
-- 也可以写成：
select department_id,job_id,avg(salary)
from hr.employees
group by department_id,grouping sets ((department_id,job_id),());

limit

• 用途:限制查询/返回的条数;
• 用法:
  • 指定初始位置: limit 初始位置,条数
  • 不指定初始位置: limit 条数
  • 由offset指定初始位置: limit 条数 offset 初始位置
• 说明:
  • 初始位置:下标(从0开始),表示从该位置开始返回;
  • 条数:表示返回的记录数

参考:https://blog.csdn.net/qq_33801641/article/details/107196036