大資料進階之路——Spark SQL日志分析-有解無憂

文章目錄

- - 基本方案
  - 資料處理流程
  - 資料清洗
  - 二次清洗
  - 視頻訪問
  - 按照省份
  - 按照流量
  - 優化
  - 資料可視化
  - echarts

基本方案

用戶行為日志：用戶每次訪問網站時所有的行為資料（訪問、瀏覽、搜索、點擊…）
用戶行為軌跡、流量日志

日志資料內容：

1）訪問的系統屬性：作業系統、瀏覽器等等
2）訪問特征：點擊的url、從哪個url跳轉過來的(referer)、頁面上的停留時間等
3）訪問資訊：session_id、訪問ip(訪問城市)等

2013-05-19 13:00:00     http://www.taobao.com/17/?tracker_u=1624169&type=1      B58W48U4WKZCJ5D1T3Z9ZY88RU7QA7B1        http://hao.360.cn/      1.196.34.243

資料處理流程

1）資料采集
Flume： web日志寫入到HDFS
2）資料清洗
臟資料
Spark、Hive、MapReduce 或者是其他的一些分布式計算框架
清洗完之后的資料可以存放在HDFS(Hive/Spark SQL)
3）資料處理
按照我們的需要進行相應業務的統計和分析
Spark、Hive、MapReduce 或者是其他的一些分布式計算框架
4）處理結果入庫
結果可以存放到RDBMS、NoSQL
5）資料的可視化
通過圖形化展示的方式展現出來：餅圖、柱狀圖、地圖、折線圖
ECharts、HUE、Zeppelin

資料清洗

首先通過debug 找到分割后各個欄位的對應的

報錯

java.io.IOException: Could not locate executable null\bin\winutils.exe in the Hadoop binaries.

執行第一步資料清洗時候，資料能列印出來，但是不能寫入本地檔案，這是因為本地沒有hadoop偽分布式系統

裝一個插件即可

https://hiszm.lanzous.com/iWyqmhrgk0f

下載上述插件，然后，新建目錄并且放入到目錄里面
C:\Data\hadoop\bin

然后再系統環境變數添加
HADOOP_HOME
C:\Data\hadoop

package org.sparksql

import org.apache.spark.sql.SparkSession

object SparkFormatApp {

  def main(args: Array[String]): Unit = {

    //SparkSession是spark的入口類
    val spark = SparkSession.builder().appName("SparkFormatApp")
                .master("local[2]").getOrCreate()
    val access = spark.sparkContext.textFile("10000_access.log")

    //access.take(10).foreach(println)

    access.map(line=>{
      val splits = line.split(" ")
      val ip = splits(0)
      val time = splits(3) + " " + splits(4)
      val traffic = splits(9)
      val url =  splits(11).replace("\"","")
     //(ip,DateUtils.parse(time),traffic,traffic,url)
      DateUtils.parse(time) + "\t" + url + "\t" + traffic + "\t" + ip
    }).saveAsTextFile("output")

    //.take(10).foreach(println)
    //.saveAsTextFile("output")

    spark.stop()

  }
}

一般的日志處理方式，我們是需要進行磁區的，
按照日志中的訪問時間進行相應的磁區，比如：d,h,m5(每5分鐘一個磁區)

二次清洗

輸入：訪問時間、訪問URL、耗費的流量、訪問IP地址資訊
輸出：URL、cmsType(video/article)、cmsId(編號)、流量、ip、城市資訊、訪問時間、天

package org.sparksql

import org.apache.spark.sql.Row
import org.apache.spark.sql.types.{LongType, StringType, StructField, StructType}

//訪問日志工具轉換類
object AccessConvertUtils {

  val struct=StructType(
    Array(
      StructField("url",StringType),
      StructField("cmsType",StringType),
      StructField("cmsId",LongType),
      StructField("traffic",LongType),
      StructField("ip",StringType),
      StructField("city",StringType),
      StructField("time",StringType),
      StructField("day",StringType)
    )
  )

//根據輸入的每一行資訊轉化成輸出的樣式
  def parseLog(log:String)={
    try{
      val splits=log.split("\t")
      val url =splits(1)
      val traffic = splits(2).toLong
      val ip = splits(3)

      val domain="http://www.imooc.com/"
      val cms=url.substring(url.indexOf(domain) + domain.length)
      val cmsTypeId = cms.split("/")
      var cmsType = ""
      var cmsId = 0l
      if(cmsTypeId.length > 1){
        cmsType = cmsTypeId(0)
        cmsId = cmsTypeId(1).toLong
      }

      val city = IpUtils.getCity(ip)
      val time = splits(0)
      val day =  time.substring(0,10).replaceAll("-","")
      Row(url,cmsType,cmsId,traffic,ip,city,time,day)
    }catch {
      case e : Exception => Row(0)
    }
  }
}

IP=>省份

使用github上已有的開源專案
1）git clone https://github.com/wzhe06/ipdatabase.git

2）編譯下載的專案：mvn clean package -DskipTests

3）安裝jar包到自己的maven倉庫

mvn install:install-file -Dfile=C:\Data\ipdatabase\target\ipdatabase-1.0-SNAPSHOT.jar -DgroupId=com.ggstar -DartifactId=ipdatabase -Dversion=1.0 -Dpackaging=jar

拷貝相關檔案不然會報錯

java.io.FileNotFoundException: file:/Users/rocky/maven_repos/com/ggstar/ipdatabase/1.0/ipdatabase-1.0.jar!/ipRegion.xlsx (No such file or directory)

測驗

package org.sparksql

import org.apache.spark.sql.SparkSession

object SparkCleanApp {

  def main(args: Array[String]): Unit = {
    //SparkSession是spark的入口類
    val spark = SparkSession.builder().appName("SparkFormatApp")
      .master("local[2]").getOrCreate()
    val accessRDD = spark.sparkContext.textFile("access.log")

    //accessRDD.take(10).foreach(println)

    val accessDF = spark.createDataFrame(accessRDD.map(x=>AccessConvertUtils.parseLog(x)),AccessConvertUtils.struct)

    accessDF.printSchema()
    accessDF.show()

    spark.stop
  }


}

root
 |-- url: string (nullable = true)
 |-- cmsType: string (nullable = true)
 |-- cmsId: long (nullable = true)
 |-- traffic: long (nullable = true)
 |-- ip: string (nullable = true)
 |-- city: string (nullable = true)
 |-- time: string (nullable = true)
 |-- day: string (nullable = true)




+--------------------+-------+-----+-------+---------------+----+-------------------+--------+
|                 url|cmsType|cmsId|traffic|             ip|city|               time|     day|
+--------------------+-------+-----+-------+---------------+----+-------------------+--------+
|http://www.imooc....|  video| 4500|    304|  218.75.35.226| 浙江省|2017-05-11 14:09:14|20170511|
|http://www.imooc....|  video|14623|     69| 202.96.134.133| 廣東省|2017-05-11 15:25:05|20170511|
|http://www.imooc....|article|17894|    115| 202.96.134.133| 廣東省|2017-05-11 07:50:01|20170511|
|http://www.imooc....|article|17896|    804|  218.75.35.226| 浙江省|2017-05-11 02:46:43|20170511|
|http://www.imooc....|article|17893|    893|222.129.235.182| 北京市|2017-05-11 09:30:25|20170511|
|http://www.imooc....|article|17891|    407|  218.75.35.226| 浙江省|2017-05-11 08:07:35|20170511|
|http://www.imooc....|article|17897|     78| 202.96.134.133| 廣東省|2017-05-11 19:08:13|20170511|
|http://www.imooc....|article|17894|    658|222.129.235.182| 北京市|2017-05-11 04:18:47|20170511|
|http://www.imooc....|article|17893|    161|   58.32.19.255| 上海市|2017-05-11 01:25:21|20170511|
|http://www.imooc....|article|17895|    701|    218.22.9.56| 安徽省|2017-05-11 13:37:22|20170511|
|http://www.imooc....|article|17892|    986|  218.75.35.226| 浙江省|2017-05-11 05:53:47|20170511|
|http://www.imooc....|  video|14540|    987|   58.32.19.255| 上海市|2017-05-11 18:44:56|20170511|
|http://www.imooc....|article|17892|    610|  218.75.35.226| 浙江省|2017-05-11 17:48:51|20170511|
|http://www.imooc....|article|17893|      0|    218.22.9.56| 安徽省|2017-05-11 16:20:03|20170511|
|http://www.imooc....|article|17891|    262|   58.32.19.255| 上海市|2017-05-11 00:38:01|20170511|
|http://www.imooc....|  video| 4600|    465|  218.75.35.226| 浙江省|2017-05-11 17:38:16|20170511|
|http://www.imooc....|  video| 4600|    833|222.129.235.182| 北京市|2017-05-11 07:11:36|20170511|
|http://www.imooc....|article|17895|    320|222.129.235.182| 北京市|2017-05-11 19:25:04|20170511|
|http://www.imooc....|article|17898|    460| 202.96.134.133| 廣東省|2017-05-11 15:14:28|20170511|
|http://www.imooc....|article|17899|    389|222.129.235.182| 北京市|2017-05-11 02:43:15|20170511|
+--------------------+-------+-----+-------+---------------+----+-------------------+--------+

調優點：

控制檔案輸出的大小： coalesce
磁區欄位的資料型別調整：spark.sql.sources.partitionColumnTypeInference.enabled
批量插入資料庫資料，提交使用batch操作

package org.sparksql

import org.apache.spark.sql.{DataFrame, SparkSession}
import org.apache.spark.sql.functions._
object TopNApp {
  //最受歡迎
  def videoAccessTopN(spark: SparkSession, accessDF: DataFrame) = {
    import spark.implicits._
    val videoTopNDF = accessDF.filter($"day"==="20170511"&& $"cmsType" === "video")
      .groupBy("day","cmsId").agg(count("cmsId")
      .as("times")).orderBy($"times".desc)
    videoTopNDF.show()

    accessDF.createOrReplaceTempView("access_log")
    val videoTopNDF1 = spark.sql("select day,cmsId,count(1) as times from access_log where day='20170511' and cmsType = 'video' group by day,cmsId order by times desc")

    videoTopNDF1.show()


  }

  def main(args: Array[String]): Unit = {
    //SparkSession是spark的入口類
    val spark = SparkSession.builder().appName("SparkFormatApp")
      .config("spark.sql.sources.partitionColumnTypeInference.enabled","false")
      .master("local[2]").getOrCreate()

    val accessDF= spark.read.format("parquet").load("output2/")
    accessDF.printSchema()
    accessDF.show(false)

    videoAccessTopN(spark,accessDF)
    spark.stop()
  }




}








+--------+-----+------+
|     day|cmsId| times|
+--------+-----+------+
|20170511|14540|111027|
|20170511| 4000| 55734|
|20170511|14704| 55701|
|20170511|14390| 55683|
|20170511|14623| 55621|
|20170511| 4600| 55501|
|20170511| 4500| 55366|
|20170511|14322| 55102|
+--------+-----+------+

視頻訪問

package org.sparksql

import java.sql.{Connection, DriverManager, PreparedStatement}

object MySqlUtils {

  def getConnection() ={

//    if (!conn.isClosed) System.out.println("已連接上資料庫!")
//    else System.out.println("沒有連接到資料庫!")
    DriverManager.getConnection("jdbc:mysql://localhost:3306/imooc_user?user=root&password=root")
  }
//釋放資料庫連接資源
  def release(connection:Connection,pstmt:PreparedStatement): Unit ={
    try{
      if(pstmt != null){
        pstmt.close()
      }

    }catch{
      case e:Exception => e.printStackTrace()

    }finally {
      if(connection!=null){
        connection.close()
      }
    }
  }

  def main(args: Array[String]): Unit = {
    println(getConnection())

  }

}


create table day_video_access_topn_stat (
day varchar(8) not null,
cms_id bigint(10) not null,
times bigint(10) not null,
primary key (day, cms_id)
);

package org.sparksql

import java.sql.{Connection, PreparedStatement}

import scala.collection.mutable.ListBuffer

object StatisticsDAO {
  def insertDayVideoAccessTopN(list:ListBuffer[DayVideoAccessStatistics]): Unit ={

    var connection:Connection = null
    var pstmt:PreparedStatement = null

    try{

      connection= MySqlUtils.getConnection()
      //取消自動提交

      connection.setAutoCommit(false)

      val sql = "insert into day_video_access_topn_stat(day,cms_id,times) value (? ,? ,? )"
      pstmt = connection.prepareStatement(sql)

      for(i<-list){
        pstmt.setString(1,i.day)
        pstmt.setLong(2,i.cmsId)
        pstmt.setLong(3,i.times)

        pstmt.addBatch()
      }

        pstmt.executeBatch()//批量處理

      //手動提交
      connection.commit()

    }catch {
      case e:Exception=>e.printStackTrace()
    }finally {
      MySqlUtils.release(connection,pstmt)
    }



  }

}

package org.sparksql

import org.apache.spark.sql.{DataFrame, SparkSession}
import org.apache.spark.sql.functions._

import scala.collection.mutable.ListBuffer
object TopNApp {
  //最受歡迎
  def videoAccessTopN(spark: SparkSession, accessDF: DataFrame) = {
    import spark.implicits._
    val videoTopNDF = accessDF.filter($"day"==="20170511"&& $"cmsType" === "video")
      .groupBy("day","cmsId").agg(count("cmsId")
      .as("times")).orderBy($"times".desc)


    videoTopNDF.show()

      try{
        videoTopNDF.foreachPartition(partitionOfRecords =>{
          val list = new ListBuffer[DayVideoAccessStatistics]

          partitionOfRecords.foreach(info =>{
            val day = info.getAs[String]("day")
            val cmsId = info.getAs[Long]("cmsId")
            val times = info.getAs[Long]("times")

            list.append(DayVideoAccessStatistics(day,cmsId,times))

          })
          StatisticsDAO.insertDayVideoAccessTopN(list)
        })


      }catch {
        case e:Exception =>e.printStackTrace()
      }

  }

java.sql.SQLException: No value specified for parameter 2

檢查插入引數和型別是否一直

按照省份

create table day_video_city_access_topn_stat (
day varchar(8) not null,
cms_id bigint(10) not null,
city varchar(20) not null,
times bigint(10) not null,
times_rank int not null,
primary key (day, cms_id, city)
);

  def cityAccessTopN(spark: SparkSession, accessDF: DataFrame) = {
    import spark.implicits._
    val cityTopNDF = accessDF.filter($"day"==="20170511"&& $"cmsType" === "video")
      .groupBy("day","city","cmsId").agg(count("cmsId")
      .as("times")).orderBy($"times".desc)
    cityTopNDF.show()

   val top3DF =  cityTopNDF.select(
      cityTopNDF("day"),
      cityTopNDF("city"),
      cityTopNDF("cmsId"),
      cityTopNDF("times"),
      row_number().over(Window.partitionBy(cityTopNDF("city"))
        .orderBy(cityTopNDF("times").desc)).as("times_rank")
    ).filter("times_rank <=3")//.show()


    try{
      top3DF.foreachPartition(partitionOfRecords =>{
        val list = new ListBuffer[DayCityAccessStatistics]
        partitionOfRecords.foreach(info =>{
          val day = info.getAs[String]("day")
          val cmsId = info.getAs[Long]("cmsId")
          val city = info.getAs[String]("city")
          val times = info.getAs[Long]("times")
          val timesRank = info.getAs[Int]("times_rank")
          list.append(DayCityAccessStatistics(day,cmsId,city,times,timesRank))
        })
        StatisticsDAO.insertCityVideoAccessTopN(list)
      })
    }catch {
      case e:Exception =>e.printStackTrace()
    }
  }

按照流量

create table day_video_traffics_topn_stat (
day varchar(8) not null,
cms_id bigint(10) not null,
traffics bigint(20) not null,
primary key (day, cms_id)
);

  def trafficAccessTopN(spark: SparkSession, accessDF: DataFrame) = {
    import spark.implicits._
    val trafficTopNDF = accessDF.filter($"day"==="20170511"&& $"cmsType" === "video")
      .groupBy("day","cmsId").agg(sum("traffic").as("traffics"))
     .orderBy($"traffics".desc)


    trafficTopNDF.show()

    try{
      trafficTopNDF.foreachPartition(partitionOfRecords =>{
        val list = new ListBuffer[DayTrafficAccessStatistics]

        partitionOfRecords.foreach(info =>{
          val day = info.getAs[String]("day")
          val cmsId = info.getAs[Long]("cmsId")
          val traffics = info.getAs[Long]("traffics")

          list.append(DayTrafficAccessStatistics(day,cmsId,traffics))

        })
        StatisticsDAO.insertTrafficVideoAccessTopN(list)
      })
    }catch {
      case e:Exception =>e.printStackTrace()
    }

  }

優化

每次更新洗掉前面的資料

  def deleteData(day:String)={
    val tables= Array("day_video_traffics_topn_stat","day_video_city_access_topn_stat","day_video_access_topn_stat")
    var connection:Connection = null
    var pstmt:PreparedStatement = null
    try{

      connection = MySqlUtils.getConnection()
      for(table<- tables){
        val deleteSQL = s"delete from $table where day = ?"
        pstmt = connection.prepareStatement(deleteSQL)
        pstmt.setString(1,day)
        pstmt.executeUpdate()
      }

    }catch {
      case e:Exception => e.printStackTrace()
    }finally {
      MySqlUtils.release(connection, pstmt)
    }
  }

資料可視化

資料可視化：一副圖片最偉大的價值莫過于它能夠使得我們實際看到的比我們期望看到的內容更加豐富

常見的可視化框架
1）echarts
2）highcharts
3）D3.js
4）HUE
5）Zeppelin

echarts

package org.sparkSQL.Utils;
import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.PreparedStatement;
import java.sql.ResultSet;
public class MySqlUtils {

    private  static final String USERNAME = "root";
    private  static final String PASSWORD = "root";
    private  static final String DRIVERCLASS = "com.mysql.jdbc.Driver";
    private  static final String URL = "jdbc:mysql://localhost:3306/imooc_user";

    public static Connection getConnection(){

        Connection connection =  null;
        try {
            Class.forName(DRIVERCLASS);
            connection  = DriverManager.getConnection(URL,USERNAME,PASSWORD);
        }catch (Exception e){
            e.printStackTrace();
        }

        return connection;


    }



    public static void  release(Connection connection, PreparedStatement pstmt , ResultSet rs){
        if(rs != null){
            try{
                rs.close();
            }catch (Exception e){
                e.printStackTrace();
            }

        }
        if(connection != null){
            try{
                connection.close();
            }catch (Exception e){
                e.printStackTrace();
            }

        }

        if(pstmt != null){
            try{
                pstmt.close();
            }catch (Exception e){
                e.printStackTrace();
            }

        }

    }

    public static void main(String[] args) {
        System.out.println(getConnection());
    }


}

<!DOCTYPE html>
<html>
<head>
    <meta charset="utf-8">
    <title>ECharts</title>
    <!-- 引入 echarts.js -->
    <script src="https://cdn.bootcss.com/echarts/3.7.1/echarts.min.js"></script>
    <script src="https://s3.pstatp.com/cdn/expire-1-M/jquery/3.1.1/jquery.min.js"></script>
</head>
<body>
<!-- 為ECharts準備一個具備大小（寬高）的Dom -->
<div id="main" style="width: 600px;height:400px;"></div>
<script type="text/javascript">


    // 基于準備好的dom，初始化echarts實體
    var myChart = echarts.init(document.getElementById('main'));

    // 指定圖表的配置項和資料
    var option = {
        title: {
            text: '最受歡迎的TOPN',
            subtext: '測驗',
            left: 'center'
        },
        tooltip: {
            trigger: 'item',
            formatter: '{a} <br/>{b} : {c} ({d}%)'
        },
        legend: {
            orient: 'vertical',
            left: 'left',
            data: ['直接訪問', '郵件營銷', '聯盟廣告', '視頻廣告', '搜索引擎']
        },
        series: [
            {
                name: '訪問次數',
                type: 'pie',
                radius: '55%',
                center: ['50%', '60%'],
                data: (function(){
                    var courses= [];
                    $.ajax({
                        type:"GET",
                        url:"stat?day=20170511",
                        dataType:'json',
                        async:false,
                        success:function (result){
                            for(var i=0;i<result.length;i++){
                                courses.push({"value":result[i].value,"name":result[i].name})
                            }

                        }
                    })
                    return courses;

                })(),
                emphasis: {
                    itemStyle: {
                        shadowBlur: 10,
                        shadowOffsetX: 0,
                        shadowColor: 'rgba(0, 0, 0, 0.5)'
                    }
                }
            }
        ]
    };


    // 使用剛指定的配置項和資料顯示圖表，
    myChart.setOption(option);
</script>
</body>
</html>

[hadoop@hadoop001 software]$ tar -zxvf zeppelin-0.7.1-bin-all.tgz -C ~/app/

[hadoop@hadoop001 bin]$ ./zeppelin-daemon.sh start
Log dir doesn't exist, create /home/hadoop/app/zeppelin-0.7.1-bin-all/logs
Pid dir doesn't exist, create /home/hadoop/app/zeppelin-0.7.1-bin-all/run
Zeppelin start

轉載請註明出處，本文鏈接：https://www.uj5u.com/ruanti/306030.html

標籤：其他

上一篇：線性方程直接求法——高斯消元法(一）

下一篇：如何給撲克洗牌才能更公平？