hive 数据库操作
hive 支持大部分的 sql 语法，因此熟悉 sql 可以很方便的上手 hive 操作
hero
-- hero
-- 载入数据
create table hero(line string);
load data inpath '/hero.txt' to table hero;
-- 提取出 英雄名字，胜负，玩家id 信息
create table info as select
regexp_extract(line, '^"(\w+),(\d)"\s#\s(\d+)$', 1) as name,
regexp_extract(line, '^"(\w+),(\d)"\s#\s(\d+)$', 2) as is_win,
regexp_extract(line, '^"(\w+),(\d)"\s#\s(\d+)$', 3) as user_id
from hero;
-- 计算 用户id，该用户总场次，和 总胜率
select
user_id,
count(1) as total_count,
sum(cast(is_win as int))/count(1) as win_rate
from info group by user_id with rollup;
-- 计算英雄，该英雄场次及胜率
select
name,
count(1) as total_count,
sum(cast(is_win as int))/count(1) as win_rate
from info group by name with rollup;
-- 英雄胜率 top3
select
name,
count(1) as total_count,
sum(cast(is_win as int))/count(1) as win_rate
from info group by name order by win_rate desc limit 3;
-- 计算 用户id， 英雄名字， 该用户该英雄出场次数 和 胜率
select
user_id,
name,
count(1) as total_count,
sum(cast(is_win as int))/count(1) as rate
from info group by user_id, name;
gamelog
-- gamelog
-- 提取 gamelog 各字段信息
create table info as select
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 1) as id,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 2) as device,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 3) as device2,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 4) as online_date,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 5) as online_time,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 6) as offline_date,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 7) as offline_time,
regexp_extract(line, '^\s*([\\w-]+)\\s+((Android|iOS)\\s+[\\d.]+)\\s+([\\d-]+)T([\\d:]+)\\s+([\\d-]+)T([\\d:]+)\\s+(\\d+)\\s*$', 8) as gametime
from gamelog;
-- 安卓端 iOS端 用户数量（注意 id 去重）
select
device2,
count(distinct id) as count
from info group by device2;
/*
Android 9948
iOS 23318
*/
-- 安卓端 iOS端 各版本 用户数量（注意 id 去重）
select
device,
count(distinct id) as count
from info group by device;
/*
Android 4.3 2037
Android 4.4 1883
Android 5.0 2011
Android 6.0 1969
Android 7.0 2048
iOS 10.1.1 3294
iOS 10.2 3249
iOS 10.3.2 3352
iOS 11.0 3298
iOS 11.1 3399
iOS 11.2 3290
iOS 11.2.5 3436
*/
-- 安卓端 iOS端 各版本 登录次数，平均在线时长
select
device,
count(1) as login_count,
avg(cast(gametime as int)) as aver_time
from info group by device;
/*
Android 4.3 20303 24598.16017337339
Android 4.4 18840 24282.822399150744
Android 5.0 20202 24659.246064746065
Android 6.0 19402 24436.93505824142
Android 7.0 20519 24462.314440274866
iOS 10.1.1 32800 24391.159786585366
iOS 10.2 32661 24450.062643519796
iOS 10.3.2 33852 24518.667316554413
iOS 11.0 33304 24433.191718712467
iOS 11.1 33859 24444.414719867687
iOS 11.2 33525 24549.974645786726
iOS 11.2.5 34619 24442.049019324648
*/
-- 每天各时段登陆的人数（去重）
select online_date, stage, count(distinct id) as count from (
select id, online_date, (
case
when time between 0 and 6 then 'dawn'
when time between 7 and 12 then 'morning'
when time between 13 and 15 then 'mid'
when time between 16 and 19 then 'afternoon'
when time between 20 and 24 then 'night'
else 'error'
end
) as stage from (
select id, online_date, cast(split(online_time, ":")[0] as int) as time from info
)a
)b group by online_date, stage;
-- 7-12 早 morning
-- 13-15 中 mid
-- 16-19 下 afternoon
-- 20-24 晚 night
-- 1-6 凌晨 dawn
结果示例
hive> select online_date, stage, count(distinct id) as count from (
> select id, online_date, (
> case
> when time between 0 and 6 then 'dawn'
> when time between 7 and 12 then 'morning'
> when time between 13 and 15 then 'mid'
> when time between 16 and 19 then 'afternoon'
> when time between 20 and 24 then 'night'
> else 'error'
> end
> ) as stage from (
> select id, online_date, cast(split(online_time, ":")[0] as int) as time from info
> )a
> )b group by online_date, stage;
WARNING: Hive-on-MR is deprecated in Hive 2 and may not be available in the future versions. Consider using a different execution engine (i.e. spark, tez) or using Hive 1.X releases.
Query ID = root_20180728223233_d170698a-a243-410c-a24e-b3c12f6aaf84
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks not specified. Estimated from input data size: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=
In order to set a constant number of reducers:
set mapreduce.job.reduces=
Starting Job = job_1532741886556_0012, Tracking URL = http://master:8088/proxy/application_1532741886556_0012/
Kill Command = /opt/SoftWare/hadoop-2.7.3/bin/hadoop job -kill job_1532741886556_0012
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 1
2018-07-28 22:34:58,189 Stage-1 map = 0%, reduce = 0%
2018-07-28 22:35:59,527 Stage-1 map = 0%, reduce = 0%
2018-07-28 22:36:38,337 Stage-1 map = 67%, reduce = 0%, Cumulative CPU 52.62 sec
2018-07-28 22:36:45,013 Stage-1 map = 100%, reduce = 0%, Cumulative CPU 58.61 sec
2018-07-28 22:37:16,082 Stage-1 map = 100%, reduce = 45%, Cumulative CPU 62.77 sec
2018-07-28 22:37:19,870 Stage-1 map = 100%, reduce = 67%, Cumulative CPU 64.69 sec
2018-07-28 22:37:35,500 Stage-1 map = 100%, reduce = 68%, Cumulative CPU 75.08 sec
2018-07-28 22:37:38,039 Stage-1 map = 100%, reduce = 72%, Cumulative CPU 77.26 sec
2018-07-28 22:37:42,296 Stage-1 map = 100%, reduce = 74%, Cumulative CPU 79.44 sec
2018-07-28 22:37:45,238 Stage-1 map = 100%, reduce = 75%, Cumulative CPU 81.63 sec
2018-07-28 22:37:49,095 Stage-1 map = 100%, reduce = 77%, Cumulative CPU 83.79 sec
2018-07-28 22:37:51,691 Stage-1 map = 100%, reduce = 80%, Cumulative CPU 85.79 sec
2018-07-28 22:37:55,689 Stage-1 map = 100%, reduce = 83%, Cumulative CPU 88.48 sec
2018-07-28 22:37:58,208 Stage-1 map = 100%, reduce = 90%, Cumulative CPU 91.65 sec
2018-07-28 22:38:01,819 Stage-1 map = 100%, reduce = 99%, Cumulative CPU 94.74 sec
2018-07-28 22:38:03,065 Stage-1 map = 100%, reduce = 100%, Cumulative CPU 96.58 sec
MapReduce Total cumulative CPU time: 1 minutes 36 seconds 580 msec
Ended Job = job_1532741886556_0012
MapReduce Jobs Launched:
Stage-Stage-1: Map: 1 Reduce: 1 Cumulative CPU: 96.58 sec HDFS Read: 32762295 HDFS Write: 1308 SUCCESS
Total MapReduce CPU Time Spent: 1 minutes 36 seconds 580 msec
OK
2017-01-01 afternoon 3177
2017-01-01 dawn 11076
2017-01-01 mid 2567
2017-01-01 morning 2676
2017-01-01 night 5880
2017-01-02 afternoon 4085
2017-01-02 dawn 13930
2017-01-02 mid 3173
2017-01-02 morning 3285
2017-01-02 night 7412
2017-01-03 afternoon 4478
2017-01-03 dawn 15393
2017-01-03 mid 3423
2017-01-03 morning 3690
2017-01-03 night 8315
2017-01-04 afternoon 4666
2017-01-04 dawn 16371
2017-01-04 mid 3699
2017-01-04 morning 3795
2017-01-04 night 8766
2017-01-05 afternoon 5827
2017-01-05 dawn 20144
2017-01-05 mid 4427
2017-01-05 morning 4762
2017-01-05 night 10822
2017-01-06 afternoon 7507
2017-01-06 dawn 26158
2017-01-06 mid 5838
2017-01-06 morning 6249
2017-01-06 night 13990
2017-01-07 afternoon 7083
2017-01-07 dawn 24763
2017-01-07 mid 5405
2017-01-07 morning 5803
2017-01-07 night 13380
Time taken: 332.701 seconds, Fetched: 35 row(s)
hive>