在Haoop集群提交指令
yarn jar BigDataTraining.jar jike.hadoop.phoneDataProcess.PhoneDataProcessor xieyue/HTTP_20130313143750.dat xieyue/out任务执行结果
执行结果
代码路径：https://github.com/DepInjoy/geektime/tree/main/BigDataTraining/src/main/java/jike/hadoop/hbaseAction
提交指令
yarn jar BigDataTraining.jar jike.hadoop.hbaseAction.StudentInfoProcessor
查询语句：
SELECT t_user.age, AVG(t_rating.rate) AS avgrate
FROM hive_sql_test1.t_rating AS t_rating
INNER JOIN hive_sql_test1.t_user AS t_user
ON t_user.userid=t_rating.userid
WHERE t_rating.movieid='2116'
GROUP BY t_user.age;
查询语句：
SELECT collect_set(t_user.sex)[0] AS sex, t_movie.moviename,
AVG(t_rating.rate) AS avgrate, COUNT(*) AS total
FROM hive_sql_test1.t_user AS t_user
INNER JOIN hive_sql_test1.t_rating AS t_rating
ON t_user.userid=t_rating.userid
INNER JOIN hive_sql_test1.t_movie AS t_movie
ON t_movie.movieid=t_rating.movieid
WHERE t_user.sex='M'
GROUP BY t_movie.moviename
HAVING total>50
ORDER BY avgrate DESC
LIMIT 10;
查询语句：
CREATE VIEW IF NOT EXISTS t_f_view AS
SELECT t_user.userid AS userid,
t_movie.moviename AS moviename,
t_rating.rate AS rate
From hive_sql_test1.t_movie AS t_movie
INNER JOIN hive_sql_test1.t_rating AS t_rating
ON t_movie.movieid=t_rating.movieid
INNER JOIN hive_sql_test1.t_user AS t_user
ON t_rating.userid=t_user.userid
WHERE t_user.sex='F';
SELECT moviename, avgrate
FROM (
SELECT moviename, AVG(t_rating.rate) AS avgrate
FROM hive_sql_test1.t_movie AS t_movie
INNER JOIN hive_sql_test1.t_rating AS t_rating
ON t_movie.movieid=t_rating.movieid
GROUP BY t_movie.moviename
) AS t_static_info
LEFT SEMI JOIN (
SELECT t_f_view.rate AS rate, t_f_view.moviename AS moviename,
top1_user.userid
FROM t_f_view, (
SELECT userid, COUNT(*) AS total
FROM t_f_view
GROUP BY userid
ORDER BY total DESC
LIMIT 1) top1_user
WHERE t_f_view.userid=top1_user.userid
ORDER BY rate DESC
LIMIT 10
) t_top1user_top10rate
ON t_static_info.moviename=t_top1user_top10rate.moviename
ORDER BY avgrate DESC;
主要用途是基数统计，例如 -1 统计页面的UV -2 统计访问量(IP数) -3 统计在线用户数 -4 统计每天搜索不同词条的个数 -5 统计注册 IP 数 -6 统计在线用户数 -7 统计用户每天搜索不同词条的个数
select approx_distinct(user_id) from user_visits;
select distinct count(user_id) from user_visits;每分钟（小时）计算在五分钟（小时）内每个账号的平均交易金额
git clone https://github.com/apache/flink-playgrounds.git功能实现: geektime\BigDataTraining\src\main\java\jike\flink\SpendReport.java
配置hosts:
47.101.206.149 jikehadoop01
47.101.216.12 jikehadoop02
47.101.204.23 jikehadoop03
47.101.202.85 jikehadoop04
47.101.72.185 jikehadoop05
139.196.15.153 jikehadoop06
106.15.39.86 jikehadoop07
139.196.162.22 jikehadoop08