hive 向用户推荐好友收藏的电影

需求

向用户 user_id = 1 推荐其关注的人喜欢的音乐

有如下三个表 1、用户关注表

user_id	follow_id
1	7
1	5
1	4
1	3

2、collect_movies

user_id	movice_id
1	1001
1	1004
1	1002
3	1001
3	1005
3	1006
4	1004
4	1007
5	1002
5	1006
7	1005

3、电影码表

movice_id	movice_name
1001	笑傲镜湖
1002	倚天屠龙记
1003	神雕侠侣
1004	孔雀翎
1005	流星蝴蝶剑
1006	萍踪侠影
1007	侠客行

数据准备

use default;
set mapreduce.map.memory.mb=81920;
set mapreduce.reduce.memory.mb=81920;
create table user_follow_0207
(
    user_id     int,
    follower_id int
) row format delimited
    fields terminated by '\t';

insert into  user_follow_0207 values (1, 7);
insert into  user_follow_0207 values (1, 4);
insert into  user_follow_0207 values (1, 5);
insert into  user_follow_0207 values (1, 3);

create table collect_movies_0207
(
    user_id  int,
    movice_id int
) row format delimited
    fields terminated by '\t';

insert into collect_movies_0207 values (1  ,1001);
insert into collect_movies_0207 values (1  ,1004);
insert into collect_movies_0207 values (1  ,1002);
insert into collect_movies_0207 values (3  ,1001);
insert into collect_movies_0207 values (3  ,1005);
insert into collect_movies_0207 values (3  ,1006);
insert into collect_movies_0207 values (4  ,1004);
insert into collect_movies_0207 values (4  ,1007);
insert into collect_movies_0207 values (5  ,1002);
insert into collect_movies_0207 values (5  ,1006);
insert into collect_movies_0207 values (7  ,1005);
 create table movies
(
    id         int,
    muvie_name varchar(50)
) row format delimited
    fields terminated by '\t';

insert into movies values (1001, "笑傲镜湖");
insert into movies values (1002    ,"倚天屠龙记");
insert into movies values (1003    ,"神雕侠侣");
insert into movies values (1004    ,"孔雀翎");
insert into movies values (1005,"流星蝴蝶剑");
insert into movies values (1006    ,"萍踪侠影");
insert into movies values (1007,"侠客行");

需求处理

 1、首先将收藏表和电影码表进行关联

select t.user_id,t.movice_id,t1.muvie_name 
from collect_movies_0207 t
         left join movies t1
                   on t.movice_id = t1.id;

user_id	movice_id	muvie_name
1	1001	笑傲镜湖
1	1004	孔雀翎
7	1005	流星蝴蝶剑
1	1002	倚天屠龙记
3	1001	笑傲镜湖
3	1005	流星蝴蝶剑
3	1006	萍踪侠影
4	1004	孔雀翎
4	1007	侠客行
5	1002	倚天屠龙记
5	1006	萍踪侠影

2、拿用户关注表和第一步处理的结果集进行关联 

select *
from user_follow_0207 t
         left join
     (select t.user_id, t.movice_id, t1.music_name
      from collect_movies_0207 t
               left join movies t1
                         on t.movice_id = t1.id) t1
     on t.follower_id = t1.user_id

user_id	follow_id	user_id	movice_id	muvie_name
1	7	7	1005	流星蝴蝶剑
1	4	4	1004	孔雀翎
1	4	4	1007	侠客行
1	5	5	1002	倚天屠龙记
1	5	5	1006	萍踪侠影
1	3	3	1001	笑傲镜湖
1	3	3	1006	萍踪侠影
1	3	3	1005	流星蝴蝶剑

 3、根据步骤2结果集，向用户ID为1推荐好友，采用collect_set()函数进行聚合，通过collect_set()函数也可以将重复的电影进行去重 

select t.user_id,
concat_ws("-", collect_set(t1.muvie_name)) muvie_names
from (select * from user_follow_0207 where user_id = 1) t
         left join
     (select t.user_id, t.movice_id, u.muvie_name
      from collect_movies_0207 t
               left join movies u on t.movice_id = u.id) t1
     on t.follower_id = t1.user_id
group by t.user_id;

user_id	muvie_names
1	笑傲镜湖,萍踪侠影,倚天屠龙记,侠客行,孔雀翎,流星蝴蝶剑

小结

这个需求主要是考察了 collect_set函数的用法。