Spark - COVID-19 案例实践使用 ScalaJavaPython 三种语言演示

Posted 2022-12-15 小毕超

tags:

篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了Spark - COVID-19 案例实践使用 ScalaJavaPython 三种语言演示相关的知识，希望对你有一定的参考价值。

一、COVID-19 案例

在前面的文章中，有使用 MapReduce 实践 COVID-19 新冠肺炎案例，本篇基于 Spark 对该数据集进行分析，下面是 MapReduce 案例的文章地址：

https://blog.csdn.net/qq_43692950/article/details/127475811

数据格式如下所示：

date（日期）,county（县）,state（州）,fips（县编码code）,cases（累计确诊病例）,deaths（累计死亡病例）
2021-01-28,Pike,Alabama,01109,2704,35
2021-01-28,Randolph,Alabama,01111,1505,37
2021-01-28,Russell,Alabama,01113,3675,16
2021-01-28,Shelby,Alabama,01117,19878,141
2021-01-28,St. Clair,Alabama,01115,8047,147
2021-01-28,Sumter,Alabama,01119,925,28
2021-01-28,Talladega,Alabama,01121,6711,114
2021-01-28,Tallapoosa,Alabama,01123,3258,112
2021-01-28,Tuscaloosa,Alabama,01125,22083,283
2021-01-28,Walker,Alabama,01127,6105,185
2021-01-28,Washington,Alabama,01129,1454,27

数据集下载：

https://download.csdn.net/download/qq_43692950/86805389

二、计算各个州的累积cases、deaths

思路：

读取数据集
转换键值对，key：州，value：各个县的 cases、deaths 值
根据 key 聚合，相加 cases、deaths 值
根据州排序输出

Scala：

object CovidSum 

  case class CovidVO(cases: Long, deaths: Long) extends java.io.Serializable

  def main(args: Array[String]): Unit = 
    val conf = new SparkConf().setAppName("spark").setMaster("local[*]")
    val sc = new SparkContext(conf)
    sc.setLogLevel("WARN")

    val t1 = sc.textFile("D:/test/input/")
    t1.filter(StringUtils.isNotBlank)
      .map(s => 
        val line = s.split(",")
        if (line.size >= 6) 
          val state = line(2)
          val cases = line(4).toLong
          val deaths = line(5).toLong
          (state, CovidVO(cases, deaths))
         else 
          null
        
      ).filter(Objects.nonNull)
      .reduceByKey((v1, v2) => CovidVO(v1.cases + v2.cases, v1.deaths + v2.deaths))
      .sortByKey(ascending = true, 1)
      .foreach(t => 
        println(t._1 + "  " + t._2.cases + "  " + t._2.deaths)
      )

Java：

public class CvoidSum 

    @Data
    @AllArgsConstructor
    @NoArgsConstructor
    static class CovidVO implements Serializable 
        private Long cases;
        private Long deaths;
    

    public static void main(String[] args) 
        SparkConf conf = new SparkConf().setAppName("spark").setMaster("local[*]");
        JavaSparkContext sc = new JavaSparkContext(conf);
        sc.setLogLevel("WARN");
        JavaRDD<String> t1 = sc.textFile("D:/test/input/");
        t1.filter(StringUtils::isNotBlank)
                .mapToPair(s -> 
                    List<String> line = Arrays.asList(s.split(","));
                    if (line.size() >= 6) 
                        String state = line.get(2);
                        Long cases = Long.parseLong(line.get(4));
                        Long deaths = Long.parseLong(line.get(5));
                        return new Tuple2<>(state, new CovidVO(cases, deaths));
                    
                    return null;
                ).filter(Objects::nonNull)
                .reduceByKey((v1, v2) -> new CovidVO(v1.getCases() + v2.getCases(), v1.getDeaths() + v2.getDeaths()))
                .sortByKey(true, 1)
                .foreach(t -> 
                    System.out.println(t._1 + "  " + t._2.getCases() + "  " + t._2.getDeaths());
                );

Python：

from pyspark import SparkConf, SparkContext
import findspark

if __name__ == '__main__':
    findspark.init()
    conf = SparkConf().setAppName('spark').setMaster('local[*]')
    sc = SparkContext(conf=conf)
    sc.setLogLevel("WARN")

    t1 = sc.textFile("D:/test/input/")

    def mapT(s):
        line = s.split(",")
        if (len(line) >= 6):
            state = line[2]
            casesStr = line[4]
            deathsStr = line[5]
            cases = int(casesStr if (casesStr and casesStr != '') else '0')
            deaths = int(deathsStr if (deathsStr and deathsStr != '') else '0')
            return (state, (cases, deaths))
        else:
            return None

    t1.filter(lambda s: s and s != '') \\
        .map(mapT) \\
        .filter(lambda s: s != None) \\
        .reduceByKey(lambda v1, v2: (v1[0] + v2[0], v1[1] + v2[1])) \\
        .sortByKey(ascending=True, numPartitions=1) \\
        .foreach(lambda t: print(t[0] + "  " + str(t[1][0]) + "  " + str(t[1][1])))

统计结果：

三、计算各个州的累积cases、deaths，并根据 deaths 降序排列

思路：

读取数据集
转换键值对，key：州，value：各个县的 cases、deaths 值
根据 key 聚合，相加 cases、deaths 值
根据 value 中的 deaths 降序排列输出

Scala:

object CovidSumSortDeaths 
  def main(args: Array[String]): Unit = 
    val conf = new SparkConf().setAppName("spark").setMaster("local[*]")
    val sc = new SparkContext(conf)
    sc.setLogLevel("WARN")

    val t1 = sc.textFile("D:/test/input/")
    t1.filter(StringUtils.isNotBlank)
      .map(s => 
        val line = s.split(",")
        if (line.size >= 6) 
          val state = line(2)
          val cases = line(4).toLong
          val deaths = line(5).toLong
          (state, CovidVO(cases, deaths))
         else 
          null
        
      ).filter(Objects.nonNull)
      .reduceByKey((v1, v2) => CovidVO(v1.cases + v2.cases, v1.deaths + v2.deaths))
      .sortBy(_._2.deaths, ascending = false, 1)
      .foreach(t => 
        println(t._1 + "  " + t._2.cases + "  " + t._2.deaths)
      )

Java:

public class CvoidSumSortDeaths 

    @Data
    @AllArgsConstructor
    @NoArgsConstructor
    static class CovidVO implements Serializable 
        private Long cases;
        private Long deaths;
    

    public static void main(String[] args) 
        SparkConf conf = new SparkConf().setAppName("spark").setMaster("local[*]");
        JavaSparkContext sc = new JavaSparkContext(conf);
        sc.setLogLevel("WARN");
        JavaRDD<String> t1 = sc.textFile("D:/test/input/");
        t1.filter(StringUtils::isNotBlank)
                .mapToPair(s -> 
                    List<String> line = Arrays.asList(s.split(","));
                    if (line.size() >= 6) 
                        String state = line.get(2);
                        Long cases = Long.parseLong(line.get(4));
                        Long deaths = Long.parseLong(line.get(5));
                        return new Tuple2<>(state, new CovidVO(cases, deaths));
                    
                    return null;
                ).filter(Objects::nonNull)
                .reduceByKey((v1, v2) -> new CovidVO(v1.getCases() + v2.getCases(), v1.getDeaths() + v2.getDeaths()))
                .map(t -> new Tuple2<String, CovidVO>(t._1, t._2))
                .sortBy(t -> t._2.getDeaths(), false, 1)
                .foreach(t -> 
                    System.out.println(t._1 + "  " + t._2.getCases() + "  " + t._2.getDeaths());
                );

Python:

from pyspark import SparkConf, SparkContext
import findspark

if __name__ == '__main__':
    findspark.init()
    conf = SparkConf().setAppName('spark').setMaster('local[*]')
    sc = SparkContext(conf=conf)
    sc.setLogLevel("WARN")

    t1 = sc.textFile("D:/test/input/")

    def mapT(s):
        line = s.split(",")
        if (len(line) >= 6):
            state = line[2]
            casesStr = line[4]
            deathsStr = line[5]
            cases = int(casesStr if (casesStr and casesStr != '') else '0')
            deaths = int(deathsStr if (deathsStr and deathsStr != '') else '0')
            return (state, (cases, deaths))
        else:
            return None

    t1.filter(lambda s: s and s != '') \\
        .map(mapT) \\
        .filter(lambda s: s != None) \\
        .reduceByKey(lambda v1, v2: (v1[0] + v2[0], v1[1] + v2[1])) \\
        .sortBy(lambda t: t[1][1], ascending=False, numPartitions=1) \\
        .foreach(lambda t: print(t[0] + "  " + str(t[1][0]) + "  " + str(t[1][1])))

统计结果：

四、计算各个州根据 deaths 排列Top3的县

思路：

读取数据集
转换键值对，key：州，value：各个县以及cases、deaths 值
根据 key 聚合，相加 cases、deaths 值
声明一个长度为 3 的数组，来实时计算存储 top3 的值，从而减少全量排序的 Shuffle 占用内存过大问题
先计算出每个分区 top3 的值
对各个分区 top3 的值进行比较求取最终 top3 的值
根据州排序输出

Scala：

object CovidTopDeaths 

  case class CovidVO(county: String, cases: Long, deaths: Long) extends java.io.Serializable

  def main(args: Array[String]): Unit = 
    val conf = new SparkConf().setAppName("spark").setMaster("local[*]")
    val sc = new SparkContext(conf)
    sc.setLogLevel("WARN")

    val t1 = sc.textFile("D:/test/input/")
    t1.filter(StringUtils.isNotBlank)
      .map(s => 
        val line = s.split(",")
        if (line.size >= 6) 
          val state = line(2)
          val county = line(1)
          val cases = line(4).toLong
          val deaths = line(5).toLong
          (state, CovidVO(county, cases, deaths))
         else 
          null
        
      ).filter(Objects.nonNull)
      .aggregateByKey(new Array[CovidVO](3))(
        (zeroValue: Array[CovidVO], currentValue: CovidVO) => 
          comparison(zeroValue, currentValue)
        ,
        (topP1: Array[CovidVO], topP2: Array[CovidVO]) => 
          var top = topP2
          for (topP1Value <- topP1) 
            top = comparison(top, topP1Value)
          
          top
        
      ).sortByKey(ascending = true, 1)
      .foreach(t => 
        for (topValue <- t._2) 
          if (Objects.nonNull(topValue)) println(t._1 + "  " + topValue)
        
      )

  <以上是关于Spark - COVID-19 案例实践使用 ScalaJavaPython 三种语言演示的主要内容，如果未能解决你的问题，请参考以下文章 
 Hadoop3 - MapReduce COVID-19 案例实践
 Hadoop3 - MapReduce COVID-19 案例实践
 单细胞文献-8 单细胞RNA测序揭示外周和肺对COVID-19的不同免疫反应
 数学建模基于matlab SIR模型新冠病毒COVID-19估计含Matlab源码 2042期
 数学建模基于matlab SIR模型新冠病毒COVID-19估计含Matlab源码 2042期
 observablehq 美国 COVID-19 确诊数曲线