presto高级用法(grouping、grouping sets)

drop database if exists db_test cascade;

create database db_test;

create table db_test.tb_student(
    name string,
    score   int,
    city    string,
    sex string,
    hobby string
)
row format delimited fields terminated by '\t';

load data local inpath '/test/student.txt' into table db_test.tb_student;

select * from db_test.tb_student;

student.txt数据

张三 10 北京男喝酒

李四 20 北京男抽烟

王五 30 北京女烫头

赵六 40 上海男抽烟

麻七 50 上海女烫头

在presto中计算

分解式

按照城市分组统计人数

sql 复制代码

select city,count(1) as cnt from hive.db_test.tb_student group by city;

按照性别分组统计人数

sql 复制代码

select hobby,count(1) as cnt from hive.db_test.tb_student group by hobby;

按照爱好分组统计人数

sql 复制代码

select hobby,count(1) as cnt from hive.db_test.tb_student group by hobby;

按照城市和性别分组统计人数

sql 复制代码

select city, sex, count(1) as cnt from hive.db_test.tb_student group by city, sex;

按照城市和爱好分组统计人数

sql 复制代码

select city, hobby, count(1) as cnt from hive.db_test.tb_student group by city, hobby;

按照性别和爱好分组统计人数

sql 复制代码

select sex, hobby, count(1) as cnt from hive.db_test.tb_student group by sex, hobby;

按照城市和性别还有爱好分组统计人数

sql 复制代码

select city, sex, hobby, count(1) as cnt from hive.db_test.tb_student group by city, sex, hobby;

统计人数

sql 复制代码

select count(1) as cnt from hive.db_test.tb_student group by ();

合并式

sql 复制代码

with t1 as (
    select city, null as sex, null as hobby, count(1) as cnt, 1 as o from hive.db_test.tb_student group by city
    union all
    select null as city, sex, null as hobby, count(1) as cnt, 2 as o from hive.db_test.tb_student group by sex
    union all
    select null, null, hobby,count(1) as cnt, 3 as o from hive.db_test.tb_student group by hobby
    union all
    select city, sex, null, count(1) as cnt, 4 as o from hive.db_test.tb_student group by city, sex
    union all
    select city, null, hobby, count(1) as cnt, 5 as o from hive.db_test.tb_student group by city, hobby
    union all
    select null, sex, hobby, count(1) as cnt, 6 as o from hive.db_test.tb_student group by sex, hobby
    union all
    select city, sex, hobby, count(1) as cnt, 7 as o from hive.db_test.tb_student group by city, sex, hobby
    union all
    select null, null, null, count(1) as cnt, 8 as o from hive.db_test.tb_student group by ()
)
select * from t1
order by o, city, sex, hobby
;

presto使用grouping

sql 复制代码

select
    city,
    sex,
    count(1) as cnt,
    grouping(city, sex) as g
from hive.db_test.tb_student
group by city, sex
;

presto使用grouping sets

sql 复制代码

select
    city,
    sex,
    hobby,
    count(1) as cnt,
    grouping(city, sex, hobby)
from hive.db_test.tb_student
group by grouping sets (city, sex, hobby)
;

sql 复制代码

select
    city,
    sex,
    hobby,
    count(1) as cnt,
    grouping(city, sex, hobby)
from hive.db_test.tb_student
group by grouping sets (city, sex, hobby, (city, sex), (city, hobby), (sex, hobby), (city, sex, hobby), ())
;

sql 复制代码

select
    city,
    sex,
    hobby,
    count(1) as cnt,
    case
        when grouping(city, sex, hobby)=3 then 1
        when grouping(city, sex, hobby)=5 then 2
        when grouping(city, sex, hobby)=6 then 3
        when grouping(city, sex, hobby)=1 then 4
        when grouping(city, sex, hobby)=2 then 5
        when grouping(city, sex, hobby)=4 then 6
        when grouping(city, sex, hobby)=0 then 7
        when grouping(city, sex, hobby)=7 then 8
        else 100
    end as o
from hive.db_test.tb_student
group by grouping sets (city, sex, hobby, (city, sex), (city, hobby), (sex, hobby), (city, sex, hobby), ())
order by o, city, sex, hobby
;

grouping作用例子展示

sql 复制代码

with t1 as (
    select '北京' as city, '男' as sex
    union all
    select '北京' as city, '男' as sex
    union all
    select '北京' as city, '女' as sex
    union all
    select '北京' as city, null as sex
)
select
    city,
    sex,
    count(1) as cnt
from t1
group by grouping sets (city, (city, sex))

复制代码

问题:
    city=北京, sex=null, cnt=4
    city=北京, sex=null, cnt=1
    为什么 city 和 sex 的值一样, 但是结果不同?
原因:
    一个null 表示跟这一列没有关系
    另一个null 表示 这一列的值 为null, 根据 列值统计的结果
    怎么区分
解决方案:
    grouping(city, sex)
        0,0     两个都有关
        0,1     只跟city有关
        1,0     只跟sex有关
        1,1     都这两列都无关

sql 复制代码

with t1 as (
    select '北京' as city, '男' as sex
    union all
    select '北京' as city, '男' as sex
    union all
    select '北京' as city, '女' as sex
    union all
    select '北京' as city, null as sex
)
select
    city,
    sex,
    count(1) as cnt,
    grouping(city, sex) g
from t1
group by grouping sets (city, (city, sex))

sql 复制代码

select
    city,
    sex,
    hobby,
    count(1) as cnt,
    case
        when grouping(city, sex, hobby)=3 then 1
        when grouping(city, sex, hobby)=5 then 2
        when grouping(city, sex, hobby)=6 then 3
        when grouping(city, sex, hobby)=1 then 4
        when grouping(city, sex, hobby)=2 then 5
        when grouping(city, sex, hobby)=4 then 6
        when grouping(city, sex, hobby)=0 then 7
        when grouping(city, sex, hobby)=7 then 8
        else 100
    end as o
from hive.db_test.tb_student
group by grouping sets (city, sex, hobby, (city, sex), (city, hobby), (sex, hobby), (city, sex, hobby), ())
order by o, city, sex, hobby

高级用法: cube

sql 复制代码

select
    city,
    sex,
    hobby,
    count(1) as cnt,
    case
        when grouping(city, sex, hobby)=3 then 1
        when grouping(city, sex, hobby)=5 then 2
        when grouping(city, sex, hobby)=6 then 3
        when grouping(city, sex, hobby)=1 then 4
        when grouping(city, sex, hobby)=2 then 5
        when grouping(city, sex, hobby)=4 then 6
        when grouping(city, sex, hobby)=0 then 7
        when grouping(city, sex, hobby)=7 then 8
        else 100
    end as o
from hive.db_test.tb_student
group by cube(city, sex, hobby)
order by o, city, sex, hobby

rollup 用法

sql 复制代码

select
    city,
    sex,
    hobby,
    count(1) as cnt,
    case
        when grouping(city, sex, hobby)=3 then 1
        when grouping(city, sex, hobby)=5 then 2
        when grouping(city, sex, hobby)=6 then 3
        when grouping(city, sex, hobby)=1 then 4
        when grouping(city, sex, hobby)=2 then 5
        when grouping(city, sex, hobby)=4 then 6
        when grouping(city, sex, hobby)=0 then 7
        when grouping(city, sex, hobby)=7 then 8
        else 100
    end as o
from hive.db_test.tb_student
group by rollup(city, sex, hobby)
order by o, city, sex, hobby
;

总结：

presto时间函数：

date()类型表示年月日

timestamp类型表示年月日时分秒

eg：timestamp（'2024-08-18 22:13:10','%Y-%m-%d %H%i%s'）

date_add(unit, value,timestamp)

grouping sets()相当于一个集合都能根据括号里的内容分组查询到相应的数据

grouping 根据8421码 0表示与该列有关系1表示无关通过计算数值查看与列之间分组的关系

cube(city, sex, hobby) 等价于 grouping sets (city, sex, hobby, (city, sex), (city, hobby), (sex, hobby), (city, sex, hobby), ())

rollup (city, sex, name) 等价于 grouping set((city, sex, name), (city, sex), city, ())

presto高级用法(grouping、grouping sets)

准备工作：

在hive中建表

在presto中计算

分解式

按照城市分组 统计人数

按照性别分组 统计人数

按照爱好分组 统计人数

按照城市和性别分组 统计人数

按照城市和爱好分组 统计人数

按照性别和爱好分组 统计人数

按照城市和性别还有爱好分组 统计人数

统计人数

合并式

presto使用grouping

presto使用grouping sets

grouping作用例子展示

高级用法: cube

rollup 用法

按照城市分组统计人数

按照性别分组统计人数

按照爱好分组统计人数

按照城市和性别分组统计人数

按照城市和爱好分组统计人数

按照性别和爱好分组统计人数

按照城市和性别还有爱好分组统计人数