6 模块开发——ETL

该项目的数据分析过程在hadoop集群上实现，主要应用hive数据仓库工具，因此，采集并经过预处理后的数据，需要加载到hive数据仓库中，以进行后续的挖掘分析。

6.1创建原始数据表

--在hive仓库中建贴源数据表

drop table if exists ods_weblog_origin;

create table ods_weblog_origin(

valid string,

remote_addr string,

remote_user string,

time_local string,

request string,

status string,

body_bytes_sent string,

http_referer string,

http_user_agent string)

partitioned by (datestr string)

row format delimited

fields terminated by '\001';

点击流模型pageviews表

drop table if exists ods_click_pageviews;

create table ods_click_pageviews(

Session string,

remote_addr string,

time_local string,

request string,

visit_step string,

page_staylong string,

http_referer string,

http_user_agent string,

body_bytes_sent string,

status string)

partitioned by (datestr string)

row format delimited

fields terminated by '\001';

时间维表创建

drop table dim_time if exists ods_click_pageviews;

create table dim_time(

year string,

month string,

day string,

hour string)

row format delimited

fields terminated by ',';

6.2导入数据

导入清洗结果数据到贴源数据表ods_weblog_origin

load data inpath '/weblog/preprocessed/16-02-24-16/' overwrite into table ods_weblog_origin partition(datestr='2013-09-18');

0: jdbc:hive2://localhost:10000> show partitions ods_weblog_origin;

+-------------------+--+

| partition |

+-------------------+--+

| timestr=20151203 |

+-------------------+--+

0: jdbc:hive2://localhost:10000> select count(*) from ods_origin_weblog;

+--------+--+

| _c0 |

+--------+--+

| 11347 |

+--------+--+

导入点击流模型pageviews数据到ods_click_pageviews表

0: jdbc:hive2://hdp-node-01:10000> load data inpath '/weblog/clickstream/pageviews' overwrite into table ods_click_pageviews partition(datestr='2013-09-18');

0: jdbc:hive2://hdp-node-01:10000> select count(1) from ods_click_pageviews;

+------+--+

| _c0 |

+------+--+

| 66 |

+------+--+

导入点击流模型visit数据到ods_click_visit表

6.3 生成ODS层明细宽表

6.3.1 需求概述

整个数据分析的过程是按照数据仓库的层次分层进行的，总体来说，是从ODS原始数据中整理出一些中间表（比如，为后续分析方便，将原始数据中的时间、url等非结构化数据作结构化抽取，将各种字段信息进行细化，形成明细表），然后再在中间表的基础之上统计出各种指标数据

6.3.2 ETL实现

建表——明细表 (源：ods_weblog_origin) （目标：ods_weblog_detail）

drop table ods_weblog_detail;

create table ods_weblog_detail(

valid string, --有效标识

remote_addr string, --来源IP

remote_user string, --用户标识

time_local string, --访问完整时间

daystr string, --访问日期

timestr string, --访问时间

month string, --访问月

day string, --访问日

hour string, --访问时

request string, --请求的url

status string, --响应码

body_bytes_sent string, --传输字节数

http_referer string, --来源url[dht1]

ref_host string, --来源的host

ref_path string, --来源的路径

ref_query string, --来源参数query

ref_query_id string, --来源参数query的值

http_user_agent string --客户终端标识

)

partitioned by(datestr string);

--抽取refer_url到中间表 "t_ods_tmp_referurl"

--将来访url分离出host path query query id

drop table if exists t_ods_tmp_referurl;

create table t_ ods _tmp_referurl as

SELECT a.*,b.*

FROM ods_origin_weblog a LATERAL VIEW parse_url_tuple(regexp_replace(http_referer, "\"", ""), 'HOST', 'PATH','QUERY', 'QUERY:id') b as host, path, query, query_id;

--抽取转换time_local字段到中间表明细表”t_ ods _detail”

drop table if exists t_ods_tmp_detail;

create table t_ods_tmp_detail as

select b.*,substring(time_local,0,10) as daystr,

substring(time_local,11) as tmstr,

substring(time_local,5,2) as month,

substring(time_local,8,2) as day,

substring(time_local,11,2) as hour

From t_ ods _tmp_referurl b;

以上语句可以改写成：

insert into table zs.ods_weblog_detail partition(datestr='$day_01')

select c.valid,c.remote_addr,c.remote_user,c.time_local,

substring(c.time_local,0,10) as daystr,

substring(c.time_local,12) as tmstr,

substring(c.time_local,6,2) as month,

substring(c.time_local,9,2) as day,

substring(c.time_local,11,3) as hour,

c.request,c.status,c.body_bytes_sent,c.http_referer,c.ref_host,c.ref_path,c.ref_query,c.ref_query_id,c.http_user_agent

from

(SELECT

a.valid,a.remote_addr,a.remote_user,a.time_local,

a.request,a.status,a.body_bytes_sent,a.http_referer,a.http_user_agent,b.ref_host,b.ref_path,b.ref_query,b.ref_query_id

FROM zs.ods_weblog_origin a LATERAL VIEW parse_url_tuple(regexp_replace(http_referer, "\"", ""), 'HOST', 'PATH','QUERY', 'QUERY:id') b as ref_host, ref_path, ref_query, ref_query_id) c

0: jdbc:hive2://localhost:10000> show partitions ods_weblog_detail;

+---------------------+--+

| partition |

+---------------------+--+

| dd=18%2FSep%2F2013 |

+---------------------+--+

1 row selected (0.134 seconds)

http://www.baidu.com/aapath?sousuoci=’angel’

parse_url_tuple(url,’HOST’,’PATH’,’QUERY’,’QUERY:id’)

7 模块开发——统计分析

注：每一种统计指标都可以跟各维度表进行叉乘，从而得出各个维度的统计结果

篇幅限制，叉乘的代码及注释信息详见项目工程代码文件

为了在前端展示时速度更快，每一个指标都事先算出各维度结果存入mysql

提前准备好维表数据，在hive仓库中创建相应维表，如：

时间维表：

create table v_time(year string,month string,day string,hour string)

row format delimited

fields terminated by ',';

load data local inpath '/home/hadoop/v_time.txt' into table v_time;

在实际生产中，究竟需要哪些统计指标通常由相关数据需求部门人员提出，而且会不断有新的统计需求产生，以下为网站流量分析中的一些典型指标示例。

1. PV统计

1.1 多维度统计PV总量

1. 时间维度

--计算指定的某个小时pvs

select count(*),month,day,hour from dw_click.ods_weblog_detail group by month,day,hour;

--计算该处理批次（一天）中的各小时pvs

drop table dw_pvs_hour;

create table dw_pvs_hour(month string,day string,hour string,pvs bigint) partitioned by(datestr string);

insert into table dw_pvs_hour partition(datestr='2016-03-18')

select a.month as month,a.day as day,a.hour as hour,count(1) as pvs from ods_weblog_detail a

where a.datestr='2016-03-18' group by a.month,a.day,a.hour;

或者用时间维表关联

维度：日

drop table dw_pvs_day;

create table dw_pvs_day(pvs bigint,month string,day string);

insert into table dw_pvs_day

select count(1) as pvs,a.month as month,a.day as day from dim_time a

join ods_weblog_detail b

on b.dd='18/Sep/2013' and a.month=b.month and a.day=b.day

group by a.month,a.day;

--或者，从之前算好的小时结果中统计

Insert into table dw_pvs_day

Select sum(pvs) as pvs,month,day from dw_pvs_hour group by month,day having day='18';

结果如下：

维度：月

drop table t_display_pv_month;

create table t_display_pv_month (pvs bigint,month string);

insert into table t_display_pv_month

select count(*) as pvs,a.month from t_dim_time a

join t_ods_detail_prt b on a.month=b.month group by a.month;

2. 按终端维度统计pv总量

注：探索数据中的终端类型

select distinct(http_user_agent) from ods_weblog_detail where http_user_agent like '%Mozilla%' limit 200;

终端维度：uc

drop table t_display_pv_terminal_uc;

create table t_display_pv_ terminal_uc (pvs bigint,mm string,dd string,hh string);

终端维度：chrome

drop table t_display_pv_terminal_chrome;

create table t_display_pv_ terminal_ chrome (pvs bigint,mm string,dd string,hh string);

终端维度：safari

drop table t_display_pv_terminal_safari;

create table t_display_pv_ terminal_ safari (pvs bigint,mm string,dd string,hh string);

3. 按栏目维度统计pv总量

栏目维度：job

栏目维度：news

栏目维度：bargin

栏目维度：lane

1.2 人均浏览页数

需求描述：比如，今日所有来访者，平均请求的页面数

--总页面请求数/去重总人数

drop table dw_avgpv_user_d;

create table dw_avgpv_user_d(

day string,

avgpv string);

insert into table dw_avgpv_user_d

select '2013-09-18',sum(b.pvs)/count(b.remote_addr) from

(select remote_addr,count(1) as pvs from ods_weblog_detail where datestr='2013-09-18' group by remote_addr) b;

1.3 按referer维度统计pv总量

需求：按照来源及时间维度统计PVS，并按照PV大小倒序排序

-- 按照小时粒度统计，查询结果存入：( "dw_pvs_referer_h" )

drop table dw_pvs_referer_h;

create table dw_pvs_referer_h(referer_url string,referer_host string,month string,day string,hour string,pv_referer_cnt bigint) partitioned by(datestr string);

insert into table dw_pvs_referer_h partition(datestr='2016-03-18')

select http_referer,ref_host,month,day,hour,count(1) as pv_referer_cnt

from ods_weblog_detail

group by http_referer,ref_host,month,day,hour

having ref_host is not null

order by hour asc,day asc,month asc,pv_referer_cnt desc;

按天粒度统计各来访域名的访问次数并排序

drop table dw_ref_host_visit_cnts_h;

create table dw_ref_host_visit_cnts_h(ref_host string,month string,day string,hour string,ref_host_cnts bigint) partitioned by(datestr string);

insert into table dw_ref_host_visit_cnts_h partition(datestr='2016-03-18')

select ref_host,month,day,hour,count(1) as ref_host_cnts

from ods_weblog_detail

group by ref_host,month,day,hour

having ref_host is not null

order by hour asc,day asc,month asc,ref_host_cnts desc;

注：还可以按来源地域维度、访客终端维度等计算

1.4 统计pv总量最大的来源TOPN

需求描述：按照时间维度，比如，统计一天内产生最多pvs的来源topN

需要用到row_number函数

以下语句对每个小时内的来访host次数倒序排序标号，

selectref_host,ref_host_cnts,concat(month,hour,day),

row_number() over(partition by concat(month,hour,day) order by ref_host_cnts desc) as od

from dw_ref_host_visit_cnts_h

效果如下：

根据上述row_number的功能，可编写Hql取各小时的ref_host访问次数topn

drop table dw_pvs_refhost_topn_h;

create table dw_pvs_refhost_topn_h(

hour string,

toporder string,

ref_host string,

ref_host_cnts string

) partitioned by(datestr string);

insert into table zs.dw_pvs_refhost_topn_h partition(datestr='2016-03-18')

select t.hour,t.od,t.ref_host,t.ref_host_cnts from

(select ref_host,ref_host_cnts,concat(month,day,hour) as hour,

row_number() over (partition by concat(month,day,hour) order by ref_host_cnts desc) as od

from zs.dw_ref_host_visit_cnts_h) t where od<=3;

结果如下：

注：还可以按来源地域维度、访客终端维度等计算

2. 受访分析

统计每日最热门的页面top10

drop table dw_pvs_d;

create table dw_pvs_d(day string,url string,pvs string);

insert into table dw_pvs_d

select '2013-09-18',a.request,a.request_counts from

(select request as request,count(request) as request_counts from ods_weblog_detail where datestr='2013-09-18' group by request having request is not null) a

order by a.request_counts desc limit 10;

结果如下：

注：还可继续得出各维度交叉结果

3. 访客分析

3.1 独立访客

需求描述：按照时间维度比如小时来统计独立访客及其产生的pvCnts

对于独立访客的识别，如果在原始日志中有用户标识，则根据用户标识即很好实现;

此处，由于原始日志中并没有用户标识，以访客IP来模拟，技术上是一样的，只是精确度相对较低

时间维度：时

drop table dw_user_dstc_ip_h;

create table dw_user_dstc_ip_h(

remote_addr string,

pvs bigint,

hour string);

insert into table dw_user_dstc_ip_h

select remote_addr,count(1) as pvs,concat(month,day,hour) as hour

from ods_weblog_detail

Where datestr='2013-09-18'

group by concat(month,day,hour),remote_addr;

在此结果表之上，可以进一步统计出，每小时独立访客总数，每小时请求次数topn访客等

如每小时独立访客总数：

select count(1) as dstc_ip_cnts,hour from dw_user_dstc_ip_h group by hour;

练习：

统计每小时请求次数topn的独立访客

时间维度：月

select remote_addr,count(1) as counts,month

from ods_weblog_detail

group by month,remote_addr;

时间维度：日

select remote_addr,count(1) as counts,concat(month,day) as day

from ods_weblog_detail

Where dd='18/Sep/2013'

group by concat(month,day),remote_addr;

注：还可以按来源地域维度、访客终端维度等计算

3.2 每日新访客

需求描述：将每天的新访客统计出来

实现思路：创建一个去重访客累积表，然后将每日访客对比累积表

时间维度：日

--历日去重访客累积表

drop table dw_user_dsct_history;

create table dw_user_dsct_history(

day string,

ip string

)

partitioned by(datestr string);

--每日新用户追加到累计表

drop table dw_user_dsct_history;

create table dw_user_dsct_history(

day string,

ip string

)

partitioned by(datestr string);

--每日新用户追加到累计表

insert into table dw_user_dsct_history partition(datestr='2013-09-19')

select tmp.day as day,tmp.today_addr as new_ip from

(

select today.day as day,today.remote_addr as today_addr,old.ip as old_addr

from

(select distinct remote_addr as remote_addr,"2013-09-19" as day from ods_weblog_detail where datestr="2013-09-19") today

left outer join

dw_user_dsct_history old

on today.remote_addr=old.ip

) tmp

where tmp.old_addr is null;

验证：

select count(distinct remote_addr) from ods_weblog_detail;

-- 1005

select count(1) from dw_user_dsct_history where prtflag_day='18/Sep/2013';

--845

select count(1) from dw_user_dsct_history where prtflag_day='19/Sep/2013';

--160

时间维度：月

类似日粒度算法

注：还可以按来源地域维度、访客终端维度等计算

4. Visit分析（点击流模型）

4.2 回头/单次访客统计

需求描述：查询今日所有回头访客及其访问次数

实现思路：上表中出现次数>1的访客，即回头访客；反之，则为单次访客

drop table dw_user_returning;

create table dw_user_returning(

day string,

remote_addr string,

acc_cnt string)

partitioned by (datestr string);

insert overwrite table dw_user_returning partition(datestr='2013-09-18')

select tmp.day,tmp.remote_addr,tmp.acc_cnt

from

(select '2013-09-18' as day,remote_addr,count(session) as acc_cnt from click_stream_visit group by remote_addr) tmp

where tmp.acc_cnt>1;

4.3 人均访问频次

需求：统计出每天所有用户访问网站的平均次数（visit）

总visit数/去重总用户数

select sum(pagevisits)/count(distinct remote_addr) from click_stream_visit partition(datestr='2013-09-18');

5. Visit分析另一种实现方式[dht1]

5.1 mr程序识别出访客的每次访问

a.) 首先开发MAPREDUCE程序：UserStayTime

注：代码略长，见项目工程代码

b.) 提交MAPREDUCE程序进行运算

[hadoop@hdp-node-01 ~]$ hadoop jar weblog.jar cn.itcast.bigdata.hive.mr.UserStayTime /weblog/input /weblog/stayout4

--导入hive表("t_display_access_info")中

drop table ods_access_info;

create table ods_access_info(remote_addr string,firt_req_time string,last_req_time string,stay_long string)

partitioned by(prtflag_day string)

row format delimited

fields terminated by '\t';

load data inpath '/weblog/stayout4' into table ods_access_info partition(prtflag_day='18/Sep/2013');

创建表时stay_long使用的string类型，但是在后续过程中发现还是用bigint更好，进行表修改

alter table ods_access_info change column stay_long stay_long bigint;

5.2 将mr结果导入访客访问信息表"t_display_access_info"

由于有一些访问记录是单条记录，mr程序处理处的结果给的时长是0，所以考虑给单次请求的停留时间一个默认市场30秒

drop table dw_access_info;

create table dw_access_info(remote_addr string,firt_req_time string,last_req_time string,stay_long string)

partitioned by(prtflag_day string);

insert into table dw_access_info partition(prtflag_day='19/Sep/2013')

select remote_addr,firt_req_time,last_req_time,

case stay_long

when 0 then 30000

else stay_long

end as stay_long

from ods_access_info

where prtflag_day='18/Sep/2013';

在访问信息表的基础之上，可以实现更多指标统计，如：

统计所有用户停留时间平均值，观察用户在站点停留时长的变化走势

select prtflag_day as dt,avg(stay_long) as avg_staylong

from dw_access_info group by prtflag_day;

5.3 回头/单次访客统计

注：从上一步骤得到的访问信息统计表“dw_access_info”中查询

--回头访客访问信息表 "dw_access_info_htip"

drop table dw_access_info_htip;

create table dw_access_info_htip(remote_addr string, firt_req_time string, last_req_time string, stay_long string,acc_counts string)

partitioned by(prtflag_day string);

insert into table dw_access_info_htip partition(prtflag_day='18/Sep/2013')

select b.remote_addr,b.firt_req_time,b.last_req_time,b.stay_long,a.acc_counts from

(select remote_addr,count(remote_addr) as acc_counts from dw_access_info where prtflag_day='18/Sep/2013' group by remote_addr having acc_counts>1) a

join

dw_access_info b

on a.remote_addr = b.remote_addr;

--单次访客访问信息表 "dw_access_info_dcip"

drop table dw_access_info_dcip;

create table dw_access_info_dcip(remote_addr string, firt_req_time string, last_req_time string, stay_long string,acc_counts string)

partitioned by(prtflag_day string);

insert into table dw_access_dcip partition(prtflag_day='18/Sep/2013')

select b.remote_addr,b.firt_req_time,b.last_req_time,b.stay_long,a.acc_counts from

(select remote_addr,count(remote_addr) as acc_counts from dw_access_info where prtflag_day='18/Sep/2013' group by remote_addr having acc_counts<2) a

join

dw_access_info b

on a.remote_addr = b.remote_addr;

在回头/单词访客信息表的基础之上，可以实现更多统计指标，如：

--当日回头客占比

drop table dw_htpercent_d;

create table dw_htpercent_d(day string,ht_percent float);

Insert into table dw_htpercent_d

select '18/Sep/2013',(tmp_ht.ht/tmp_all.amount)*100 from

(select count( distinct a.remote_addr) as ht from dw_access_info_htip a where prtflag_day='18/Sep/2013') tmp_ht

Join

(select count(distinct b.remote_addr) as amount from dw_access_info b where prtflag_day='18/Sep/2013') tmp_all;

5.4 人均访问频度

--总访问次数/去重总人数，从访客次数汇总表中查询

select avg(user_times.counts) as user_access_freq

from

(select remote_addr,counts from t_display_htip

union all

select remote_addr,counts from t_display_access_dcip) user_times;

--或直接从访问信息表 t_display_access_info 中查询

select avg(a.acc_cts) from

(select remote_addr,count(*) as acc_cts from dw_access_info group by remote_addr) a;

6.关键路径转化率分析——漏斗模型

转化：在一条指定的业务流程中，各个步骤的完成人数及相对上一个步骤的百分比

6.1 需求分析

6.2 模型设计

定义好业务流程中的页面标识，下例中的步骤为：

Step1、 /item%

Step2、 /category

Step3、 /order

Step4、 /index

6.3 开发实现

分步骤开发：

1、查询每一个步骤的总访问人数

create table route_numbs as

select 'step1' as step,count(distinct remote_addr) as numbs from ods_click_pageviews where request like '/item%'

union

select 'step2' as step,count(distinct remote_addr) as numbs from ods_click_pageviews where request like '/category%'

union

select 'step3' as step,count(distinct remote_addr) as numbs from ods_click_pageviews where request like '/order%'

union

select 'step4' as step,count(distinct remote_addr) as numbs from ods_click_pageviews where request like '/index%';

2、查询每一步骤相对于路径起点人数的比例

思路：利用join

select rn.step as rnstep,rn.numbs as rnnumbs,rr.step as rrstep,rr.numbs as rrnumbs from route_num rn

inner join

route_num rr

select tmp.rnstep,tmp.rnnumbs/tmp.rrnumbs as ratio

from

(

select rn.step as rnstep,rn.numbs as rnnumbs,rr.step as rrstep,rr.numbs as rrnumbs from route_num rn

inner join

route_num rr) tmp

where tmp.rrstep='step1';

3、查询每一步骤相对于上一步骤的漏出率

select tmp.rrstep as rrstep,tmp.rrnumbs/tmp.rnnumbs as ration

from

(

select rn.step as rnstep,rn.numbs as rnnumbs,rr.step as rrstep,rr.numbs as rrnumbs from route_num rn

inner join

route_num rr) tmp

where cast(substr(tmp.rnstep,5,1) as int)=cast(substr(tmp.rrstep,5,1) as int)-1

4、汇总以上两种指标

select abs.step,abs.numbs,abs.ratio as abs_ratio,rel.ratio as rel_ratio

from

(

select tmp.rnstep as step,tmp.rnnumbs as numbs,tmp.rnnumbs/tmp.rrnumbs as ratio

from

(

select rn.step as rnstep,rn.numbs as rnnumbs,rr.step as rrstep,rr.numbs as rrnumbs from route_num rn

inner join

route_num rr) tmp

where tmp.rrstep='step1'

) abs

left outer join

(

select tmp.rrstep as step,tmp.rrnumbs/tmp.rnnumbs as ratio

from

(

select rn.step as rnstep,rn.numbs as rnnumbs,rr.step as rrstep,rr.numbs as rrnumbs from route_num rn

inner join

route_num rr) tmp

where cast(substr(tmp.rnstep,5,1) as int)=cast(substr(tmp.rrstep,5,1) as int)-1

) rel

on abs.step=rel.step

8 模块开发——结果导出

报表统计结果，由sqoop从hive表中导出，示例如下，详见工程代码

sqoop export \

--connect jdbc:mysql://hdp-node-01:3306/webdb --username root --password root \

--table click_stream_visit \

--export-dir /user/hive/warehouse/dw_click.db/click_stream_visit/datestr=2013-09-18 \

--input-fields-terminated-by '\001'

9 模块开发——工作流调度

注：将整个项目的数据处理过程，从数据采集到数据分析，再到结果数据的导出，一系列的任务分割成若干个oozie的工作流，并用coordinator进行协调

工作流定义示例

Ooize配置片段示例，详见项目工程

1、日志预处理mr程序工作流定义

<workflow-app name="weblogpreprocess" xmlns="uri:oozie:workflow:0.4">

<start to="firstjob"/>

<action name="firstjob">

<map-reduce>

<job-tracker>${jobTracker}</job-tracker>

<name-node>${nameNode}</name-node>

<prepare>

<delete path="${nameNode}/${outpath}"/>

</prepare>

<configuration>

<property>

<name>mapreduce.job.map.class</name>

<value>cn.itcast.bigdata.hive.mr.WeblogPreProcess$WeblogPreProcessMapper</value>

</property>

<property>

<name>mapreduce.job.output.key.class</name>

<value>org.apache.hadoop.io.Text</value>

</property>

<property>

<name>mapreduce.job.output.value.class</name>

<value>org.apache.hadoop.io.NullWritable</value>

</property>

<property>

<name>mapreduce.input.fileinputformat.inputdir</name>

<value>${inpath}</value>

</property>

<property>

<name>mapreduce.output.fileoutputformat.outputdir</name>

<value>${outpath}</value>

</property>

<property>

<name>mapred.mapper.new-api</name>

<value>true</value>

</property>

<property>

<name>mapred.reducer.new-api</name>

<value>true</value>

</property>

</configuration>

</map-reduce>

<ok to="end"/>

<error to="kill"/>

2、数据加载etl工作流定义：

<workflow-app xmlns="uri:oozie:workflow:0.5" name="hive2-wf">

<start to="hive2-node"/>

<action name="hive2-node">

<hive2 xmlns="uri:oozie:hive2-action:0.1">

<job-tracker>${jobTracker}</job-tracker>

<name-node>${nameNode}</name-node>

<configuration>

<property>

<name>mapred.job.queue.name</name>

<value>${queueName}</value>

</property>

</configuration>

<jdbc-url>jdbc:hive2://hdp-node-01:10000</jdbc-url>

<script>script.q</script>

<param>input=/weblog/outpre2</param>

</hive2>

<ok to="end"/>

<error to="fail"/>

</action>

<kill name="fail">

<message>Hive2 (Beeline) action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>

</kill>

<end name="end"/>

</workflow-app>

3、数据加载工作流所用hive脚本：

create database if not exists dw_weblog;

use dw_weblog;

drop table if exists t_orgin_weblog;

create table t_orgin_weblog(valid string,remote_addr string,

remote_user string,

time_local string,

request string,

status string,

body_bytes_sent string,

http_referer string,

http_user_agent string)

row format delimited

fields terminated by '\001';

load data inpath '/weblog/preout' overwrite into table t_orgin_weblog;

drop table if exists t_ods_detail_tmp_referurl;

create table t_ods_detail_tmp_referurl as

SELECT a.*,b.*

FROM t_orgin_weblog a

LATERAL VIEW parse_url_tuple(regexp_replace(http_referer, "\"", ""), 'HOST', 'PATH','QUERY', 'QUERY:id') b as host, path, query, query_id;

drop table if exists t_ods_detail;

create table t_ods_detail as

select b.*,substring(time_local,0,11) as daystr,

substring(time_local,13) as tmstr,

substring(time_local,4,3) as month,

substring(time_local,0,2) as day,

substring(time_local,13,2) as hour

from t_ods_detail_tmp_referurl b;

drop table t_ods_detail_prt;

create table t_ods_detail_prt(

valid string,

remote_addr string,

remote_user string,

time_local string,

request string,

status string,

body_bytes_sent string,

http_referer string,

http_user_agent string,

host string,

path string,

query string,

query_id string,

daystr string,

tmstr string,

month string,

day string,

hour string)

partitioned by (mm string,dd string);

insert into table t_ods_detail_prt partition(mm='Sep',dd='18')

select * from t_ods_detail where daystr='18/Sep/2013';

insert into table t_ods_detail_prt partition(mm='Sep',dd='19')

select * from t_ods_detail where daystr='19/Sep/2013';

更多工作流及hql脚本定义详见项目工程

下节是单元测试，和可视化展示。

数据分析高级教程(二)

6 模块开发——ETL

6.1创建原始数据表

6.2导入数据

6.3 生成ODS层明细宽表

6.3.1 需求概述

6.3.2 ETL实现

7 模块开发——统计分析

1.1 多维度统计PV总量

1.2 人均浏览页数

1.3 按referer维度统计pv总量

1.4 统计pv总量最大的来源TOPN

2. 受访分析

3.1 独立访客

3.2 每日新访客

4. Visit分析（点击流模型）

4.2 回头/单次访客统计

4.3 人均访问频次

5. Visit分析另一种实现方式[dht1]

5.1 mr程序识别出访客的每次访问

5.2 将mr结果导入访客访问信息表"t_display_access_info"

5.3 回头/单次访客统计

5.4 人均访问频度

6.关键路径转化率分析——漏斗模型

6.1 需求分析

6.2 模型设计

6.3 开发实现

8 模块开发——结果导出

9 模块开发——工作流调度

工作流定义示例

1、日志预处理mr程序工作流定义

2、数据加载etl工作流定义：

3、数据加载工作流所用hive脚本：

热门文章

最新文章

相关课程

相关电子书

相关实验场景

热门

活动广场

任务中心

开发者评测

高校计划

乘风者计划

训练营

阿里云MVP

话题

直播

下载

镜像站

技术资料

插件

数据分析高级教程(二)

6 模块开发——ETL

6.1创建原始数据表

6.2导入数据

6.3 生成ODS层明细宽表

6.3.1 需求概述

6.3.2 ETL实现

7 模块开发——统计分析

1.1 多维度统计PV总量

1.2 人均浏览页数

1.3 按referer维度统计pv总量

1.4 统计pv总量最大的来源TOPN

2. 受访分析

3.1 独立访客

3.2 每日新访客

4. Visit分析（点击流模型）

4.2 回头/单次访客统计

4.3 人均访问频次

5. Visit分析另一种实现方式[dht1]

5.1 mr程序识别出访客的每次访问

5.2 将mr结果导入访客访问信息表"t_display_access_info"

5.3 回头/单次访客统计

5.4 人均访问频度

6.关键路径转化率分析——漏斗模型

6.1 需求分析

6.2 模型设计

6.3 开发实现

8 模块开发——结果导出

9 模块开发——工作流调度

工作流定义示例

1、日志预处理mr程序工作流定义

2、数据加载etl工作流定义：

3、数据加载工作流所用hive脚本：

热门文章

最新文章

相关课程

相关电子书

相关实验场景