spark与kafka集成进行实时 nginx代理 这种sdk埋点 原生日志实时解析 处理

日志格式
202.108.16.254^A1546795482.600^A/cntv.gif?appId=3&areaId=8213&srcContId=2535575&areaType=1&srcContName=%E5%88%87%E7%89%B9%E9%87%8C%E6%A2%85%E5%BC%80%E4%BA%8C%E5%BA%A6+%E5%8D%B0%E5%BA%A64-1%E5%A4%A7%E8%83%9C%E6%B3%B0%E5%9B%BD%E5%96%9C%E8%BF%8E%E5%BC%80%E9%97%A8%E7%BA%A2&clientChannel=vivo&clientVersion=2.7.2&contId=2535584&serverIp=172.16.42.154&menuId=8212&visitTime=20190107012442630&url=http%3A%2F%2Fm.cctv4g.com%2Fcntv%2Fresource%2Fcltv2%2FdramaDetailPage.jsp%3FcontId%3D2535575%26dataType%3D3%26stats_menuId%3D8212%26stats_areaId%3D8213%26stats_areaType%3D1%26stats_contId%3D2535584%26stats_srcContType%3D3%26stats_srcContId%3D2535575%26wdChannelName%3Dvivo%26wdVersionName%3D2.7.2%26wdClientType%3D1%26wdAppId%3D3%26wdNetType%3D4G%26uuid%3De8fb9e0c-5b59-36f6-80d7-88df323fa750&srcContType=3&appName=CCTV%E6%89%8B%E6%9C%BA%E7%94%B5%E8%A7%86++%EF%BC%88V2%EF%BC%89&netType=4G&areaName=%E6%B5%B7%E6%8A%A5&contName=%E5%88%87%E7%89%B9%E9%87%8C%E6%A2%85%E5%BC%80%E4%BA%8C%E5%BA%A6+%E5%8D%B0%E5%BA%A64-1%E5%A4%A7%E8%83%9C%E6%B3%B0%E5%9B%BD%E5%96%9C%E8%BF%8E%E5%BC%80%E9%97%A8%E7%BA%A2&sessionId=59787199A5F8278836AD26F672743C29&ua=yichengtianxia&en=e_pv&uuid=e8fb9e0c-5b59-36f6-80d7-88df323fa750&clientIp=223.104.105.169&menuName=2019%E5%B9%B4%E9%98%BF%E8%81%94%E9%85%8B%E4%BA%9A%E6%B4%B2%E6%9D%AF&clientType=1

数据视频审核记录与用户访问记录 进行了实时解析 （demo程序）
改进：硬编码改为软编码 ，解析构建成解析类，代码优化 与逻辑判断加强（多次测试还未出错）
1.离线数据后续可将转为dataframe存入hive进行仓库存储进行离线分析（spark core，sql都可以）=》存入mysql进行datav ，或者后端报表
2.实时存入mysql或者hbase进行实时展示 （前面几篇已经记载了）

import java.net.URLDecoder
import java.sql.{Connection, DriverManager}

import com.spark.common.{EventLogConstants, LoggerUtil, Test, TimeUtil}
import kafka.serializer.StringDecoder
import org.apache.hadoop.hbase.client.{ConnectionFactory, Put}
import org.apache.hadoop.hbase.util.Bytes
import org.apache.hadoop.hbase.{HBaseConfiguration, TableName}
import org.apache.log4j.Logger
import org.apache.spark.streaming.dstream.DStream
import org.apache.spark.streaming.kafka.KafkaUtils
import org.apache.spark.streaming.{Seconds, StreamingContext}
import org.apache.spark.{SparkConf, SparkContext}

import scala.collection.immutable.HashMap

object SxRlStatDemo extends Serializable {
  val logger = Logger.getLogger(classOf[LoggerUtil])
  private val serialVersionUID = -4892194648703458595L

  def main(args: Array[String]): Unit = {
    val conf = new SparkConf()
    conf.setMaster("local[2]").setAppName("sxdemo")
      .set("spark.streaming.kafka.maxRatePerPartition", "100")
      .set("spark.streaming.backpressure.enabled", "true")
    //开启被压
    val sc = SparkContext.getOrCreate(conf)
    val ssc = new StreamingContext(sc, Seconds(1))

    // 二、DStream的构建
    // kafka的Simple consumer API的连接参数， 只有两个
    // metadata.broker.list: 给定Kafka的服务器路径信息
    // auto.offset.reset：给定consumer的偏移量的值，largest表示设置为最大值，smallest表示设置为最小值(最大值&最小值指的是对应的分区中的日志数据的偏移量的值) ==> 每次启动都生效
    val kafkaParams = Map[String, String](
      "metadata.broker.list" -> "hadoop04:9092,hadoop05:9092,hadoop06:9092",
      "auto.offset.reset" -> "largest",
      "key.serializer" -> "org.apache.kafka.common.serialization.StringSerializer",
      "value.serializer" -> "org.apache.kafka.common.serialization.StringSerializer")
    //      "spark.serializer"->"org.apache.spark.serializer.KryoSerializer")
    // 给定一个由topic名称组成的set集合
    val topics = Set("topic_bc")
    val stream = KafkaUtils.createDirectStream[String, String, StringDecoder, StringDecoder](ssc, kafkaParams, topics).map(_._2)
      //      .mapog => {
      //

      //      })
      .transform(rdd => {
      rdd.map(log => {
        var map: Map[String, String] = new HashMap[String, String]
        val splits = log.split("\\^A")
        if (splits.length==3){
        val ip = splits(0).trim
        val nginxTime = TimeUtil.parseNginxServerTime2Long(splits(1).trim).toString;
        if (nginxTime != "-1") {
          nginxTime.toString
        }
        val requestStr = splits(2)
        val index = requestStr.indexOf("?")
        if (index > -1) { // 有请求参数的情况下，获取？后面的参数
          val requestBody: String = requestStr.substring(index + 1)
          var areaInfo = if (ip.nonEmpty) Test.getInfo(ip) else Array("un", "un", "un")
          val requestParames = requestBody.split("&")
          for (e <- requestParames) {
            val index = e.indexOf("=")
            if (index < 1) {
              logger.debug("次日志无法解析")
            }
            var key = ""; var value = "";
            key = e.substring(0, index)
            value = URLDecoder.decode(e.substring(index + 1), EventLogConstants.LOG_PARAM_CHARSET)
            map.+=(key -> value)
          }
          map.+=("ip" -> ip, "s_time" -> nginxTime, "country" -> areaInfo(0), "provence" -> areaInfo(1), "city" -> areaInfo(2))
        }else{ logger.debug("次日志无法解析")}
        }
        map
      })

    })
    stream.cache()
    ssc.checkpoint("checkpoint")
    val bc_personAmt = stream.filter(log => log.contains("en") && log("en") == "e_sx")
      // combine_map.get("test_101").getOrElse("不存在") //根据key取value值,如果不存在返回后面的值
      //  scala> a.get(1)
      // res0: Option[Int] = Some(2) get返回的是Option[Int]类型 不可能等于" " ==Some("e_la")
      .map(log => (log("bc_person"), 1))
      .updateStateByKey[Long]((seq: Seq[Int], state: Option[Long]) => {
      //seq:Seq[Long] 当前批次中每个相同key的value组成的Seq
      val currentValue = seq.sum
      //state:Option[Long] 代表当前批次之前的所有批次的累计的结果，val对于wordcount而言就是先前所有批次中相同单词出现的总次数
      val preValue = state.getOrElse(0L)
      Some(currentValue + preValue)
    })