大数据离线---网站日志流量分析系统(2)---数据获取和预处理

本次接上一篇,进行实际数据的获取和预处理,会有较多的代码内容

  • 数据的获取
  • 数据的预处理

数据的获取

需求

数据采集的需求广义上来说分为两大部分。

  • 是在页面采集用户的访问行为,具体开发工作:
    1、 开发页面埋点 js,采集用户访问行为
    2、 后台接受页面 js 请求记录日志
  • 是从 web 服务器上汇聚日志到 HDFS,是数据分析系统的数据采集,具体的技术实现有很多方式:
    • Shell 脚本
      优点:轻量级,开发简单
      缺点:对日志采集过程中的容错处理不便控制
    • Java 采集程序
      优点:可对采集过程实现精细控制
      缺点:开发工作量大
    • Flume 日志采集框架
      成熟的开源日志采集系统,且本身就是 hadoop 生态体系中的一员,与hadoop体系中的各种框架组件具有天生的亲和力,可扩展性强。在网站 web 流量日志分析这种场景中,对数据采集部分的可靠性、容错能力要求通常不会非常严苛,因此使用通用的 flume 日志采集框架完全可以满足需求。
      本项目即使用 flume 来实现日志采集。

Flume 日志采集系统搭建

  • 日志切割
    nginx 服务器所生成的流量日志,存放在各台 nginx 服务器上。由于 nginx 没有自动分开文件存储日志的机制, 不会自动地进行切割, 都是写在一个文件 access.log 当中的。 如果访问量很大的话,将导致日志文件容量非常大,不便于管理。 那就需要手动对这个文件进行切割。
    切割需要使用 date 命令以获得昨天的日期、使用 kill 命令向 Nginx 进程发送重新打开日志文件的信号,以及 crontab 设置执行任务周期。 我们需要每天零点将前一天的日志存为另外一个文件,这里我们就将 Nginx 位于 logs 目录中的 access.log 存为access_[yyyy-MM-dd].log 的文件。
    自动切割 shell 脚本 nginx_log.sh:
#!/bin/bash
#设置日志文件存放目录
logs_path="/usr/local/nginx/logs/"
#设置 pid 文件
pid_path="/usr/local/nginx/nginx-1.7.3/logs/nginx.pid"
#日志文件
filepath=${logs_path}"access.log"
# Source function library.
#重命名日志文件
mv ${logs_path}access.log
${logs_path}access_$(date -d '-1 day' '+%Y-%m-%d').log
#向 nginx 主进程发信号重新打开日志
kill -USR1 `cat ${pid_path}`

crontab 设置作业(每天零时零分执行):

0 0 * * * sh /usr/local/nginx/nginx_log.sh
  • Flume采集实现
    Flume 采集系统的搭建相对简单:
    1、 在个 web 服务器上部署 agent 节点,修改配置文件
    2、 启动 agent 节点,将采集到的数据汇聚到指定的 HDFS 目录中
    针对上述的 nginx 日志生成场景, 如果通过 flume( 1.6)收集,无论是 SpoolingDirectory Source 和 Exec Source 均不能满足动态实时收集的需求,在当前 flume1.7 稳定版本中,提供了一个非常好用的 TaildirSource,使用这个 source,可以监控一个目录,并且使用正则表达式匹配该目录中的文件名进行实时收集。
    核心配置如下
a1.sources = r1
a1.sources.r1.type = TAILDIR
a1.sources.r1.channels = c1
a1.sources.r1.positionFile = /var/log/flume/taildir_position.json
a1.sources.r1.filegroups = f1 f2
a1.sources.r1.filegroups.f1 = /var/log/test1/example.log
a1.sources.r1.filegroups.f2 = /var/log/test2/.*log.*

说明:

  1. filegroups:指定 filegroups,可以有多个,以空格分隔;( TailSource 可以同时监控tail 多个目录中的文件)
  2. positionFile:配置检查点文件的路径,检查点文件会以 json 格式保存已经 tail 文件的位置,解决了断点不能续传的缺陷。
  3. filegroups.: 配置每个 filegroup 的文件绝对路径,文件名可以用正则表达式匹配通过以上配置,就可以监控文件内容的增加和文件的增加。 产生和所配置的文件名正则表达式不匹配的文件,则不会被 tail。
  • 数据样本
    数据的具体内容在采集阶段其实不用太关心。代码使用的数据来源可以在这里获取
    点击跳转到数据源
8.215.204.118 - - [18/Sep/2013:06:51:35 +0000] "GET /wp-includes/js/jquery/jquery.js?ver=1.10.2 HTTP/1.1" 304 0 "http://blog.fens.me/nodejs-socketio-chat/" "Mozilla/5.0 (Windows NT 5.1; rv:23.0) Gecko/20100101 Firefox/23.0"

数据说明:
1、访客 ip 地址: 58.215.204.118
2、访客用户信息: - -
3、请求时间: [18/Sep/2013:06:51:35 +0000]
4、请求方式: GET
5、请求的 url: /wp-includes/js/jquery/jquery.js?ver=1.10.2
6、请求所用协议: HTTP/1.1
7、响应码: 304
8、返回的数据流量: 0
9、访客的来源 url: http://blog.fens.me/nodejs-socketio-chat/


数据的预处理

主要目的

过滤“不合规”数据,清洗无意义的数据格式转换和规整,根据后续的统计需求,过滤分离出各种不同主题(不同栏目 path)的基础数据。
大数据离线---网站日志流量分析系统(2)---数据获取和预处理_第1张图片
数据的处理分为三个阶段进行:
阶段一:对无意义数据的清洗,无用的数据做标记
阶段二:对相同用户的信息整合,进行页面点击流模型的梳理,参考上一篇
阶段三:点击流模型 Visits 表(按 session 聚集的页面访问信息)


  • 阶段一:对无意义数据的清洗,无用的数据做标记
    WebLogBean类的代码
package com.weblog.mr.preprocess;

import org.apache.hadoop.io.Writable;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;

public class WebLogBean implements Writable{

    private boolean valid = true;// 判断数据是否合法
    private String remote_addr;// 记录客户端的ip地址
    private String remote_user;// 记录客户端用户名称,忽略属性"-"
    private String time_local;// 记录访问时间与时区
    private String request;// 记录请求的url与http协议
    private String status;// 记录请求状态;成功是200
    private String body_bytes_sent;// 记录发送给客户端文件主体内容大小
    private String http_referer;// 用来记录从那个页面链接访问过来的
    private String http_user_agent;// 记录客户浏览器的相关信息

    public void set(boolean 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) {
        this.valid = valid;
        this.remote_addr = remote_addr;
        this.remote_user = remote_user;
        this.time_local = time_local;
        this.request = request;
        this.status = status;
        this.body_bytes_sent = body_bytes_sent;
        this.http_referer = http_referer;
        this.http_user_agent = http_user_agent;
    }

    public void setValid(boolean valid) {
        this.valid = valid;
    }

    public void setRemote_addr(String remote_addr) {
        this.remote_addr = remote_addr;
    }

    public void setRemote_user(String remote_user) {
        this.remote_user = remote_user;
    }

    public void setTime_local(String time_local) {
        this.time_local = time_local;
    }

    public void setRequest(String request) {
        this.request = request;
    }

    public void setStatus(String status) {
        this.status = status;
    }

    public void setBody_bytes_sent(String body_bytes_sent) {
        this.body_bytes_sent = body_bytes_sent;
    }

    public void setHttp_referer(String http_referer) {
        this.http_referer = http_referer;
    }

    public void setHttp_user_agent(String http_user_agent) {
        this.http_user_agent = http_user_agent;
    }

    public boolean isValid() {

        return valid;
    }

    public String getRemote_addr() {
        return remote_addr;
    }

    public String getRemote_user() {
        return remote_user;
    }

    public String getTime_local() {
        return time_local;
    }

    public String getRequest() {
        return request;
    }

    public String getStatus() {
        return status;
    }

    public String getBody_bytes_sent() {
        return body_bytes_sent;
    }

    public String getHttp_referer() {
        return http_referer;
    }

    public String getHttp_user_agent() {
        return http_user_agent;
    }

    //重新整合数据
    @Override
    public String toString() {
        StringBuilder sb = new StringBuilder();
        sb.append(this.valid);
        sb.append("\001").append(this.getRemote_addr());
        sb.append("\001").append(this.getRemote_user());
        sb.append("\001").append(this.getTime_local());
        sb.append("\001").append(this.getRequest());
        sb.append("\001").append(this.getStatus());
        sb.append("\001").append(this.getBody_bytes_sent());
        sb.append("\001").append(this.getHttp_referer());
        sb.append("\001").append(this.getHttp_user_agent());
        return sb.toString();
    }





    //重新定义WebLogBean的序列化方式,减轻无用信息的负载
    @Override
    public void write(DataOutput out) throws IOException {
        out.writeBoolean(this.valid);
        out.writeUTF(null==remote_addr?"":remote_addr);
        out.writeUTF(null==remote_user?"":remote_user);
        out.writeUTF(null==time_local?"":time_local);
        out.writeUTF(null==request?"":request);
        out.writeUTF(null==status?"":status);
        out.writeUTF(null==body_bytes_sent?"":body_bytes_sent);
        out.writeUTF(null==http_referer?"":http_referer);
        out.writeUTF(null==http_user_agent?"":http_user_agent);
    }

    @Override
    public void readFields(DataInput in) throws IOException {
        this.valid = in.readBoolean();
        this.remote_addr = in.readUTF();
        this.remote_user = in.readUTF();
        this.time_local = in.readUTF();
        this.request = in.readUTF();
        this.status = in.readUTF();
        this.body_bytes_sent = in.readUTF();
        this.http_referer = in.readUTF();
        this.http_user_agent = in.readUTF();
    }
}

处理数据清洗的方法类WebLogParser类

package com.weblog.mr.preprocess;

import java.text.ParseException;
import java.text.SimpleDateFormat;
import java.util.Locale;
import java.util.Set;

public class WebLogParser {

    static SimpleDateFormat df1= new SimpleDateFormat("dd/MMM/yyyy:HH:mm:ss", Locale.US);
    static SimpleDateFormat df2= new SimpleDateFormat("yyyy-MM-dd HH:mm:ss", Locale.US);
    //对日志文件的实际处理方法
    public static WebLogBean parser(String line){

        WebLogBean webLogBean = new WebLogBean();
        //切割
        String[] arr = line.split(" ");

        if(arr.length>11){
            //设置IP地址
            webLogBean.setRemote_addr(arr[0]);
            //设置用户信息
            webLogBean.setRemote_user(arr[1]);

            //[18/Sep/2013:06:49:18 +0000] 设置时间信息
            String time_local = formatDate(arr[3].substring(1));
            webLogBean.setTime_local(time_local);
            //HTTP/1.1
            webLogBean.setRequest(arr[6]);
            //304
            webLogBean.setStatus(arr[8]);
            //0
            webLogBean.setBody_bytes_sent(arr[9]);
            //Mozilla/4.0 (compatible;)
            webLogBean.setHttp_referer(arr[10]);

            //用户浏览器的信息较多,需要进行拼接
            //"Mozilla/5.0 (Windows NT 5.1; rv:23.0) Gecko/20100101 Firefox/23.0"
            if (arr.length > 12) {
                StringBuilder sb = new StringBuilder();
                for(int i=11;i= 400) {// 大于400,HTTP错误
                webLogBean.setValid(false);
            }

            //非法时间格式的也设置为false;
            if("-invalid_time-".equals(webLogBean.getTime_local())){
                webLogBean.setValid(false);
            }

        }else{
            return null;
        }
        return  webLogBean;

    }

    //格式化时间的方法
    private static String formatDate(String time_locale) {
        try {
            return df2.format(df1.parse(time_locale));
        } catch (ParseException e) {
            e.printStackTrace();
            return null;
        }
    }



    public static void filtStaticResource(WebLogBean bean,Set pages){
        //如果不是规定的协议请求,标记为false
        if (!pages.contains(bean.getRequest())) {
            bean.setValid(false);
        }
    }
}

执行文件处理的类WeblogPreProcess

package com.weblog.mr.preprocess;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

import java.io.IOException;
import java.util.HashSet;
import java.util.Set;

public class WeblogPreProcess {

    static class WeblogPreProcessMapper extends Mapper{

        Set pages = new HashSet();
        Text  k = new Text();
        NullWritable v = NullWritable.get();

        //从外部配置文件中加载网站的有用url分类数据 存储到maptask的内存中,用来对日志数据进行过滤
        //wp-content/themes/silesia/images/ico-twitter.png HTTP/1.1
        @Override
        protected void setup(Context context) throws IOException, InterruptedException {
            pages.add("/about");

            pages.add("/black-ip-list/");
            pages.add("/cassandra-clustor/");
            pages.add("/finance-rhive-repurchase/");
            pages.add("/hadoop-family-roadmap/");
            pages.add("/hadoop-hive-intro/");
            pages.add("/hadoop-zookeeper-intro/");
            pages.add("/hadoop-mahout-roadmap/");

        }


        @Override
        protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

            String line = value.toString();

            WebLogBean webLogBean = WebLogParser.parser(line);

            if (webLogBean !=null){
                //过滤CSS/图片/js等静态资源
                WebLogParser.filtStaticResource(webLogBean,pages);
                k.set(webLogBean.toString());

                context.write(k,v);
            }
        }

    }

    public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {

        Configuration conf = new Configuration();
        Job job = Job.getInstance(conf);

        //加载类信息
        job.setJarByClass(WeblogPreProcess.class);

        job.setMapperClass(WeblogPreProcessMapper.class);

        //加载输出的数据类型
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(NullWritable.class);

        //加载文件的位置
        FileInputFormat.setInputPaths(job, new Path("d:/weblog/input"));
        FileOutputFormat.setOutputPath(job, new Path("d:/weblog/output"));

        //设置reduce不执行
        job.setNumReduceTasks(0);

        //输出结果
        boolean res = job.waitForCompletion(true);
        System.exit(res?0:1);
    }
}

  • 阶段二:对相同用户的信息整合,进行页面点击流模型的梳理
    数据装载类PageViewsBean
package cn.itcast.mr.weblog.pageview;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;

import org.apache.hadoop.io.Writable;

public class PageViewsBean implements Writable {

	private String session;
	private String remote_addr;
	private String timestr;
	private String request;
	private int step;
	private String staylong;
	private String referal;
	private String useragent;
	private String bytes_send;
	private String status;

	public void set(String session, String remote_addr, String useragent, String timestr, String request, int step, String staylong, String referal, String bytes_send, String status) {
		this.session = session;
		this.remote_addr = remote_addr;
		this.useragent = useragent;
		this.timestr = timestr;
		this.request = request;
		this.step = step;
		this.staylong = staylong;
		this.referal = referal;
		this.bytes_send = bytes_send;
		this.status = status;
	}

	public String getSession() {
		return session;
	}

	public void setSession(String session) {
		this.session = session;
	}

	public String getRemote_addr() {
		return remote_addr;
	}

	public void setRemote_addr(String remote_addr) {
		this.remote_addr = remote_addr;
	}

	public String getTimestr() {
		return timestr;
	}

	public void setTimestr(String timestr) {
		this.timestr = timestr;
	}

	public String getRequest() {
		return request;
	}

	public void setRequest(String request) {
		this.request = request;
	}

	public int getStep() {
		return step;
	}

	public void setStep(int step) {
		this.step = step;
	}

	public String getStaylong() {
		return staylong;
	}

	public void setStaylong(String staylong) {
		this.staylong = staylong;
	}

	public String getReferal() {
		return referal;
	}

	public void setReferal(String referal) {
		this.referal = referal;
	}

	public String getUseragent() {
		return useragent;
	}

	public void setUseragent(String useragent) {
		this.useragent = useragent;
	}

	public String getBytes_send() {
		return bytes_send;
	}

	public void setBytes_send(String bytes_send) {
		this.bytes_send = bytes_send;
	}

	public String getStatus() {
		return status;
	}

	public void setStatus(String status) {
		this.status = status;
	}

	@Override
	public void readFields(DataInput in) throws IOException {
		this.session = in.readUTF();
		this.remote_addr = in.readUTF();
		this.timestr = in.readUTF();
		this.request = in.readUTF();
		this.step = in.readInt();
		this.staylong = in.readUTF();
		this.referal = in.readUTF();
		this.useragent = in.readUTF();
		this.bytes_send = in.readUTF();
		this.status = in.readUTF();

	}

	@Override
	public void write(DataOutput out) throws IOException {
		out.writeUTF(session);
		out.writeUTF(remote_addr);
		out.writeUTF(timestr);
		out.writeUTF(request);
		out.writeInt(step);
		out.writeUTF(staylong);
		out.writeUTF(referal);
		out.writeUTF(useragent);
		out.writeUTF(bytes_send);
		out.writeUTF(status);

	}

}

数据处理类

package cn.itcast.mr.weblog.pageview;

import java.io.IOException;
import java.text.ParseException;
import java.text.SimpleDateFormat;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Comparator;
import java.util.Date;
import java.util.Locale;
import java.util.UUID;

import cn.itcast.mr.weblog.preprocess.WebLogBean;
import org.apache.commons.beanutils.BeanUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;



/**
 * 
 * 将清洗之后的日志梳理出点击流pageviews模型数据
 * 
 * 输入数据是清洗过后的结果数据
 * 
 * 区分出每一次会话,给每一次visit(session)增加了session-id(随机uuid)
 * 梳理出每一次会话中所访问的每个页面(请求时间,url,停留时长,以及该页面在这次session中的序号)
 * 保留referral_url,body_bytes_send,useragent
 * 
 * 
 * @author
 * 
 */
public class ClickStreamPageView {

	static class ClickStreamMapper extends Mapper {

		Text k = new Text();
		WebLogBean v = new WebLogBean();

		@Override
		protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

			String line = value.toString();

			String[] fields = line.split("\001");
			if (fields.length < 9) return;
			//将切分出来的各字段set到weblogbean中
			v.set("true".equals(fields[0]) ? true : false, fields[1], fields[2], fields[3], fields[4], fields[5], fields[6], fields[7], fields[8]);
			//只有有效记录才进入后续处理
			if (v.isValid()) {
			        //此处用ip地址来标识用户
				k.set(v.getRemote_addr());
				context.write(k, v);
			}
		}
	}

	static class ClickStreamReducer extends Reducer {

		Text v = new Text();

		@Override
		protected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {
			ArrayList beans = new ArrayList();

			// 先将一个用户的所有访问记录中的时间拿出来排序
			try {
				for (WebLogBean bean : values) {
					WebLogBean webLogBean = new WebLogBean();
					try {
						BeanUtils.copyProperties(webLogBean, bean);
					} catch(Exception e) {
						e.printStackTrace();
					}
					beans.add(webLogBean);
				}
				//将bean按时间先后顺序排序
				Collections.sort(beans, new Comparator() {

					@Override
					public int compare(WebLogBean o1, WebLogBean o2) {
						try {
							Date d1 = toDate(o1.getTime_local());
							Date d2 = toDate(o2.getTime_local());
							if (d1 == null || d2 == null)
								return 0;
							return d1.compareTo(d2);
						} catch (Exception e) {
							e.printStackTrace();
							return 0;
						}
					}

				});

				/**
				 * 以下逻辑为:从有序bean中分辨出各次visit,并对一次visit中所访问的page按顺序标号step
				 * 核心思想:
				 * 就是比较相邻两条记录中的时间差,如果时间差<30分钟,则该两条记录属于同一个session
				 * 否则,就属于不同的session
				 * 
				 */
				
				int step = 1;
				String session = UUID.randomUUID().toString();
				for (int i = 0; i < beans.size(); i++) {
					WebLogBean bean = beans.get(i);
					// 如果仅有1条数据,则直接输出
					if (1 == beans.size()) {
						
						// 设置默认停留时长为60s
						v.set(session+"\001"+key.toString()+"\001"+bean.getRemote_user() + "\001" + bean.getTime_local() + "\001" + bean.getRequest() + "\001" + step + "\001" + (60) + "\001" + bean.getHttp_referer() + "\001" + bean.getHttp_user_agent() + "\001" + bean.getBody_bytes_sent() + "\001"
								+ bean.getStatus());
						context.write(NullWritable.get(), v);
						session = UUID.randomUUID().toString();
						break;
					}

					// 如果不止1条数据,则将第一条跳过不输出,遍历第二条时再输出
					if (i == 0) {
						continue;
					}

					// 求近两次时间差
					long timeDiff = timeDiff(toDate(bean.getTime_local()), toDate(beans.get(i - 1).getTime_local()));
					// 如果本次-上次时间差<30分钟,则输出前一次的页面访问信息
					
					if (timeDiff < 30 * 60 * 1000) {
						
						v.set(session+"\001"+key.toString()+"\001"+beans.get(i - 1).getRemote_user() + "\001" + beans.get(i - 1).getTime_local() + "\001" + beans.get(i - 1).getRequest() + "\001" + step + "\001" + (timeDiff / 1000) + "\001" + beans.get(i - 1).getHttp_referer() + "\001"
								+ beans.get(i - 1).getHttp_user_agent() + "\001" + beans.get(i - 1).getBody_bytes_sent() + "\001" + beans.get(i - 1).getStatus());
						context.write(NullWritable.get(), v);
						step++;
					} else {
						
						// 如果本次-上次时间差>30分钟,则输出前一次的页面访问信息且将step重置,以分隔为新的visit
						v.set(session+"\001"+key.toString()+"\001"+beans.get(i - 1).getRemote_user() + "\001" + beans.get(i - 1).getTime_local() + "\001" + beans.get(i - 1).getRequest() + "\001" + (step) + "\001" + (60) + "\001" + beans.get(i - 1).getHttp_referer() + "\001"
								+ beans.get(i - 1).getHttp_user_agent() + "\001" + beans.get(i - 1).getBody_bytes_sent() + "\001" + beans.get(i - 1).getStatus());
						context.write(NullWritable.get(), v);
						// 输出完上一条之后,重置step编号
						step = 1;
						session = UUID.randomUUID().toString();
					}

					// 如果此次遍历的是最后一条,则将本条直接输出
					if (i == beans.size() - 1) {
						// 设置默认停留市场为60s
						v.set(session+"\001"+key.toString()+"\001"+bean.getRemote_user() + "\001" + bean.getTime_local() + "\001" + bean.getRequest() + "\001" + step + "\001" + (60) + "\001" + bean.getHttp_referer() + "\001" + bean.getHttp_user_agent() + "\001" + bean.getBody_bytes_sent() + "\001" + bean.getStatus());
						context.write(NullWritable.get(), v);
					}
				}

			} catch (ParseException e) {
				e.printStackTrace();

			}

		}

		private String toStr(Date date) {
			SimpleDateFormat df = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss", Locale.US);
			return df.format(date);
		}

		private Date toDate(String timeStr) throws ParseException {
			SimpleDateFormat df = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss", Locale.US);
			return df.parse(timeStr);
		}

		private long timeDiff(String time1, String time2) throws ParseException {
			Date d1 = toDate(time1);
			Date d2 = toDate(time2);
			return d1.getTime() - d2.getTime();

		}

		private long timeDiff(Date time1, Date time2) throws ParseException {

			return time1.getTime() - time2.getTime();

		}

	}

	public static void main(String[] args) throws Exception {

		Configuration conf = new Configuration();
		Job job = Job.getInstance(conf);

		job.setJarByClass(ClickStreamPageView.class);

		job.setMapperClass(ClickStreamMapper.class);
		job.setReducerClass(ClickStreamReducer.class);

		job.setMapOutputKeyClass(Text.class);
		job.setMapOutputValueClass(WebLogBean.class);

		job.setOutputKeyClass(Text.class);
		job.setOutputValueClass(Text.class);

//		FileInputFormat.setInputPaths(job, new Path(args[0]));
//		FileOutputFormat.setOutputPath(job, new Path(args[1]));

		FileInputFormat.setInputPaths(job, new Path("d:/weblog/output"));
		FileOutputFormat.setOutputPath(job, new Path("d:/weblog/pageviews"));

		job.waitForCompletion(true);

	}

}

  • 阶段三:点击流模型 Visits 表(按 session 聚集的页面访问信息)
    数据装载类VisitBean
package cn.itcast.mr.weblog.vists;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;

import org.apache.hadoop.io.Writable;

public class VisitBean implements Writable {

	private String session;
	private String remote_addr;
	private String inTime;
	private String outTime;
	private String inPage;
	private String outPage;
	private String referal;
	private int pageVisits;

	public void set(String session, String remote_addr, String inTime, String outTime, String inPage, String outPage, String referal, int pageVisits) {
		this.session = session;
		this.remote_addr = remote_addr;
		this.inTime = inTime;
		this.outTime = outTime;
		this.inPage = inPage;
		this.outPage = outPage;
		this.referal = referal;
		this.pageVisits = pageVisits;
	}

	public String getSession() {
		return session;
	}

	public void setSession(String session) {
		this.session = session;
	}

	public String getRemote_addr() {
		return remote_addr;
	}

	public void setRemote_addr(String remote_addr) {
		this.remote_addr = remote_addr;
	}

	public String getInTime() {
		return inTime;
	}

	public void setInTime(String inTime) {
		this.inTime = inTime;
	}

	public String getOutTime() {
		return outTime;
	}

	public void setOutTime(String outTime) {
		this.outTime = outTime;
	}

	public String getInPage() {
		return inPage;
	}

	public void setInPage(String inPage) {
		this.inPage = inPage;
	}

	public String getOutPage() {
		return outPage;
	}

	public void setOutPage(String outPage) {
		this.outPage = outPage;
	}

	public String getReferal() {
		return referal;
	}

	public void setReferal(String referal) {
		this.referal = referal;
	}

	public int getPageVisits() {
		return pageVisits;
	}

	public void setPageVisits(int pageVisits) {
		this.pageVisits = pageVisits;
	}

	@Override
	public void readFields(DataInput in) throws IOException {
		this.session = in.readUTF();
		this.remote_addr = in.readUTF();
		this.inTime = in.readUTF();
		this.outTime = in.readUTF();
		this.inPage = in.readUTF();
		this.outPage = in.readUTF();
		this.referal = in.readUTF();
		this.pageVisits = in.readInt();

	}

	@Override
	public void write(DataOutput out) throws IOException {
		out.writeUTF(session);
		out.writeUTF(remote_addr);
		out.writeUTF(inTime);
		out.writeUTF(outTime);
		out.writeUTF(inPage);
		out.writeUTF(outPage);
		out.writeUTF(referal);
		out.writeInt(pageVisits);

	}

	@Override
	public String toString() {
		return session + "\001" + remote_addr + "\001" + inTime + "\001" + outTime + "\001" + inPage + "\001" + outPage + "\001" + referal + "\001" + pageVisits;
	}
}

数据处理类ClickStreamVisit

package cn.itcast.mr.weblog.vists;

import java.io.IOException;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Comparator;

import cn.itcast.mr.weblog.pageview.PageViewsBean;
import org.apache.commons.beanutils.BeanUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;



/**
 * 输入数据:pageviews模型结果数据
 * 从pageviews模型结果数据中进一步梳理出visit模型
 * sessionid  start-time   out-time   start-page   out-page   pagecounts  ......
 * 
 * @author
 *
 */
public class ClickStreamVisit {

	// 以session作为key,发送数据到reducer
	static class ClickStreamVisitMapper extends Mapper {

		PageViewsBean pvBean = new PageViewsBean();
		Text k = new Text();

		@Override
		protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

			String line = value.toString();
			String[] fields = line.split("\001");
			int step = Integer.parseInt(fields[5]);
			//(String session, String remote_addr, String timestr, String request, int step, String staylong, String referal, String useragent, String bytes_send, String status)
			//299d6b78-9571-4fa9-bcc2-f2567c46df3472.46.128.140-2013-09-18 07:58:50/hadoop-zookeeper-intro/160"https://www.google.com/""Mozilla/5.0"14722200
			pvBean.set(fields[0], fields[1], fields[2], fields[3],fields[4], step, fields[6], fields[7], fields[8], fields[9]);
			k.set(pvBean.getSession());
			context.write(k, pvBean);

		}

	}

	static class ClickStreamVisitReducer extends Reducer {

		@Override
		protected void reduce(Text session, Iterable pvBeans, Context context) throws IOException, InterruptedException {

			// 将pvBeans按照step排序
			ArrayList pvBeansList = new ArrayList();
			for (PageViewsBean pvBean : pvBeans) {
				PageViewsBean bean = new PageViewsBean();
				try {
					BeanUtils.copyProperties(bean, pvBean);
					pvBeansList.add(bean);
				} catch (Exception e) {
					e.printStackTrace();
				}
			}

			Collections.sort(pvBeansList, new Comparator() {

				@Override
				public int compare(PageViewsBean o1, PageViewsBean o2) {

					return o1.getStep() > o2.getStep() ? 1 : -1;
				}
			});

			// 取这次visit的首尾pageview记录,将数据放入VisitBean中
			VisitBean visitBean = new VisitBean();
			// 取visit的首记录
			visitBean.setInPage(pvBeansList.get(0).getRequest());
			visitBean.setInTime(pvBeansList.get(0).getTimestr());
			// 取visit的尾记录
			visitBean.setOutPage(pvBeansList.get(pvBeansList.size() - 1).getRequest());
			visitBean.setOutTime(pvBeansList.get(pvBeansList.size() - 1).getTimestr());
			// visit访问的页面数
			visitBean.setPageVisits(pvBeansList.size());
			// 来访者的ip
			visitBean.setRemote_addr(pvBeansList.get(0).getRemote_addr());
			// 本次visit的referal
			visitBean.setReferal(pvBeansList.get(0).getReferal());
			visitBean.setSession(session.toString());

			context.write(NullWritable.get(), visitBean);

		}

	}

	public static void main(String[] args) throws Exception {
		Configuration conf = new Configuration();
		Job job = Job.getInstance(conf);

		job.setJarByClass(ClickStreamVisit.class);

		job.setMapperClass(ClickStreamVisitMapper.class);
		job.setReducerClass(ClickStreamVisitReducer.class);

		job.setMapOutputKeyClass(Text.class);
		job.setMapOutputValueClass(PageViewsBean.class);

		job.setOutputKeyClass(NullWritable.class);
		job.setOutputValueClass(VisitBean.class);
		
		
//		FileInputFormat.setInputPaths(job, new Path(args[0]));
//		FileOutputFormat.setOutputPath(job, new Path(args[1]));
		FileInputFormat.setInputPaths(job, new Path("d:/weblog/pageviews"));
		FileOutputFormat.setOutputPath(job, new Path("d:/weblog/visitout"));
		
		boolean res = job.waitForCompletion(true);
		System.exit(res?0:1);

	}

}


上述各项操作, 在实际的生产环境中, 明显是需要周期性运行的。所以任务的工作流调度就显得特别的重要。 除了使用我们 linux 自带的定时调度外,市面上还有许多成熟的调度产品。

下一篇介绍工作流调度器。

你可能感兴趣的:(大数据,MapReduce,Hive)