Hive之SQL练习

datediff函数

regexp_replace()

 substring

 函数准备介绍：

dateddiff：求两个时间的差值

regexp_replace：替换符号

to_date：将字符串转换成时间

date_sub：求一个时间与数字之间的差值

 

round：四舍五入

floor：向下取整

ceil：向上取整

 

题目：

背景说明：
以下表记录了用户每天的蚂蚁森林低碳生活领取的记录流水。
table_name：user_low_carbon
user_id data_dt  low_carbon
用户     日期      减少碳排放（g）

蚂蚁森林植物换购表，用于记录申领环保植物所需要减少的碳排放量
table_name:  plant_carbon
plant_id plant_name low_carbon
植物编号    植物名    换购植物所需要的碳

----题目
1.蚂蚁森林植物申领统计
问题：假设2017年1月1日开始记录低碳数据（user_low_carbon），假设2017年10月1日之前满足申领条件的用户都申领了一颗p004-胡杨，
剩余的能量全部用来领取“p002-沙柳” 。
统计在10月1日累计申领“p002-沙柳” 排名前10的用户信息；以及他比后一名多领了几颗沙柳。
得到的统计结果如下表样式：
user_id  plant_count less_count(比后一名多领了几颗沙柳)
u_101    1000         100
u_088    900          400
u_103    500          …

1.创建表
create table user_low_carbon(user_id String,data_dt String,low_carbon int) row format delimited fields terminated by '\t';
create table plant_carbon(plant_id string,plant_name String,low_carbon int) row format delimited fields terminated by '\t';

2.加载数据
load data local inpath "/opt/module/data/low_carbon.txt" into table user_low_carbon;
load data local inpath "/opt/module/data/plant_carbon.txt" into table plant_carbon;

3.设置本地模式
set hive.exec.mode.local.auto=true;

(1)统计在10月1日前每个用户减少碳排放量的总和（取前11名）
select user_id,sum(low_carbon) sum_carbon
from user_low_carbon
where datediff(regexp_replace(data_dt,"/","-"),"2017-10-1")<0
group by user_id
order by sum_carbon desc
limit 11;t1

(2)取出申领胡杨的条件
select low_carbon from plant_carbon where plant_id="p004";t2

(3)取出申领沙柳的条件
select low_carbon from plant_carbon where plant_id="p002";t3

(4)求出能申领沙柳的棵数
select user_id,floor((t1.sum_carbon-t2.low_carbon)/t3.low_carbon) treeCount from t1,t2,t3;t4

select 
user_id,
floor((t1.sum_carbon-t2.low_carbon)/t3.low_carbon) 
treeCount 
from (select user_id,sum(low_carbon) sum_carbon
from user_low_carbon
where datediff(regexp_replace(data_dt,"/","-"),"2017-10-1")<0
group by user_id
order by sum_carbon desc
limit 11)t1,
(select low_carbon from plant_carbon where plant_id="p004")t2,
(select low_carbon from plant_carbon where plant_id="p002")t3;t4

u_007   66
u_013   63
u_008   53
u_005   46
u_010   45
u_014   44
u_011   39
u_009   37
u_006   32
u_002   23
u_004   22

(5)求出前一名比后一名多几棵
select user_id,treeCount,treeCount-(lead(treeCount,1) over(order by treeCount desc))
from t4
limit 10;

//先执行order by 再执行的lead

 

select user_id,treeCount,treeCount-(lead(treeCount,1) over(order by treeCount desc))
from (select 
user_id,
floor((t1.sum_carbon-t2.low_carbon)/t3.low_carbon) 
treeCount 
from (select user_id,sum(low_carbon) sum_carbon
from user_low_carbon
where datediff(regexp_replace(data_dt,"/","-"),"2017-10-1")<0
group by user_id
order by sum_carbon desc
limit 11)t1,
(select low_carbon from plant_carbon where plant_id="p004")t2,
(select low_carbon from plant_carbon where plant_id="p002")t3)t4
limit 10;

u_007   66      3
u_013   63      10
u_008   53      7
u_005   46      1
u_010   45      1
u_014   44      5
u_011   39      2
u_009   37      5
u_006   32      9
u_002   23      1

2、蚂蚁森林低碳用户排名分析
问题：查询user_low_carbon表中每日流水记录，条件为：
用户在2017年，连续三天（或以上）的天数里，
每天减少碳排放（low_carbon）都超过100g的用户低碳流水。
需要查询返回满足以上条件的user_low_carbon表中的记录流水。
例如用户u_002符合条件的记录如下，因为2017/1/2~2017/1/5连续四天的碳排放量之和都大于等于100g：
seq（key） user_id data_dt  low_carbon
xxxxx10    u_002  2017/1/2  150
xxxxx11    u_002  2017/1/2  70
xxxxx12    u_002  2017/1/3  30
xxxxx13    u_002  2017/1/3  80
xxxxx14    u_002  2017/1/4  150
xxxxx14    u_002  2017/1/5  101
备注：统计方法不限于sql、procedure、python,java等

(1)求出2017年超过100g的用户&时间
select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100;t1

(2)计算每一行数据跟前后各两行的时间差
select user_id,data_dt,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc),"/","-")) lag2,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc),"/","-")) lag1,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc),"/","-")) lead1,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc),"/","-")) lead2
from (select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1;t2

select user_id,data_dt,
lag(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc) lag2,
lag(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc) lag1,
lead(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc) lead1,
lead(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc) lead2
from (select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1;t2

select user_id,data_dt,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag2,"/","-")) lag2Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag1,"/","-")) lag1Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead1,"/","-")) lead1Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead2,"/","-")) lead2Count
from (select user_id,data_dt,
lag(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc) lag2,
lag(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc) lag1,
lead(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc) lead1,
lead(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc) lead2
from (select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1)t2;

(3)
select user_id,data_dt,
lag(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt) lag2,
lag(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt) lag1,
lead(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt) lead1,
lead(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt) lead2
from (select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1;t2

select user_id,data_dt,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag2,"/","-")) lag2Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag1,"/","-")) lag1Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead1,"/","-")) lead1Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead2,"/","-")) lead2Count
from (select user_id,data_dt,
lag(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc) lag2,
lag(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc) lag1,
lead(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt desc) lead1,
lead(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt desc) lead2
from (select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1)t2;t3

(3)求出连续3天及以上的数据
select user_id,data_dt
from t3
where (lag2=2 and lag1=1) or (lag1=1 and lead1=-1) or(lead1=-1 and lead2=-2);

select user_id,data_dt
from (select user_id,data_dt,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag2,"/","-")) lag2Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lag1,"/","-")) lag1Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead1,"/","-")) lead1Count,
datediff(regexp_replace(data_dt,"/","-"),regexp_replace(lead2,"/","-")) lead2Count
from (select user_id,data_dt,
lag(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt) lag2,
lag(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt) lag1,
lead(data_dt,1,"1970/01/01") over (partition by user_id order by data_dt) lead1,
lead(data_dt,2,"1970/01/01") over (partition by user_id order by data_dt) lead2
from (select user_id,data_dt,sum(low_carbon) sum_carbon
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1)t2)t3
where (lag2Count=2 and lag1Count=1) or (lag1Count=1 and lead1Count=-1) or(lead1Count=-1 and lead2Count=-2);

u_002   2017/1/2
u_002   2017/1/3
u_002   2017/1/4
u_002   2017/1/5
u_005   2017/1/2
u_005   2017/1/3
u_005   2017/1/4
u_008   2017/1/4
u_008   2017/1/5
u_008   2017/1/6
u_008   2017/1/7
u_009   2017/1/2
u_009   2017/1/3
u_009   2017/1/4
u_010   2017/1/4
u_010   2017/1/5
u_010   2017/1/6
u_010   2017/1/7
u_011   2017/1/1
u_011   2017/1/2
u_011   2017/1/3
u_013   2017/1/2
u_013   2017/1/3
u_013   2017/1/4
u_013   2017/1/5
u_014   2017/1/5
u_014   2017/1/6
u_014   2017/1/7

解法2：
4    1    3
5    2    3
6    3    3
7    4    3
9    5    4
10    6    4
12    7    5
13    8    5
14    9    5

(1)求出2017年超过100g的用户&时间
select user_id,data_dt,sum(low_carbon) sum_carbon,
rank() over(partition by user_id order by data_dt) rank
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100;t1

(2)求出时间与rank之间的差值
select user_id,data_dt,
date_sub(regexp_replace(data_dt,"/","-"),rank)
from t1;

select user_id,data_dt,
date_sub(regexp_replace(data_dt,"/","-"),rank) sub
from (select user_id,data_dt,sum(low_carbon) sum_carbon,
rank() over(partition by user_id order by data_dt) rank
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1;t2

(3)求出连续3天及以上的数据
select user_id,data_dt,
count(*) over(partition by user_id,sub) threeDays
from t2;t3

select user_id,data_dt
from t3
where threeDays>=3;

select user_id,data_dt
from (select user_id,data_dt,
count(*) over(partition by user_id,sub) threeDays
from (select user_id,data_dt,
date_sub(regexp_replace(data_dt,"/","-"),rank) sub
from (select user_id,data_dt,sum(low_carbon) sum_carbon,
rank() over(partition by user_id order by data_dt) rank
from user_low_carbon
where substring(data_dt,1,4)="2017"
group by user_id,data_dt
having sum_carbon>100)t1)t2)t3
where threeDays>=3
order by user_id,data_dt;

u_002   2017/1/2
u_002   2017/1/3
u_002   2017/1/4
u_002   2017/1/5
u_005   2017/1/3
u_005   2017/1/2
u_005   2017/1/4
u_008   2017/1/4
u_008   2017/1/6
u_008   2017/1/5
u_008   2017/1/7
u_009   2017/1/4
u_009   2017/1/2
u_009   2017/1/3
u_010   2017/1/4
u_010   2017/1/5
u_010   2017/1/6
u_010   2017/1/7
u_011   2017/1/2
u_011   2017/1/1
u_011   2017/1/3
u_013   2017/1/2
u_013   2017/1/3
u_013   2017/1/4
u_013   2017/1/5
u_014   2017/1/5
u_014   2017/1/6
u_014   2017/1/7

前置函数：
datediff：求两个时间的差值
regexp_replace：替换符号
to_date：将字符串转换成时间
date_sub：求一个时间与数字之间的差值

round：四舍五入
floor：向下取整
ceil：向上取整

substring

set hive.exec.mode.local.auto=true;

map(user_id_data_dt,(data_dt,sum_carbon))
grouping(user_id)

ArrayList list = new ArrayList();

reduce(user_id_data_dt,values Iterate(data_dt,sum_carbon)){

    
    date1 = 0;
    
    for(value:values){
        if(date1=="0"){
        list.add(data_dt);
        date1 = data_dt;
        }else{
        if(data_dt-date1=?1){
        list.add(data_dt);
        date1 = data_dt;
        }else{
        if(list.size>=3){
        context.write;
        list.clear;
        date1=data_dt;
        }else{
        list.clear;
        date1=data_dt;
        }
        }
        }
    }

    list.size>?3;

    list.clear;

}