1、分布式缓存缓存的是文件,当程序启动的时候把把这份文件拷贝到没给TaskManager上,当该taskManager上的task使用的时候直接调去本地副本上的文件即可。
经常用到的一个函数就是RichMapFunction和open方法。
代码:
import org.apache.commons.io.FileUtils;
import org.apache.flink.api.common.functions.RichMapFunction;
import org.apache.flink.api.java.DataSet;
import org.apache.flink.api.java.ExecutionEnvironment;
import org.apache.flink.configuration.Configuration;
import scala.Int;
import java.io.File;
import java.util.List;
public class DistributeCacheDemo {
public static void main(String[] args) throws Exception {
//创建一个离线计算的运行环境
ExecutionEnvironment env=ExecutionEnvironment.getExecutionEnvironment();
//注册一个本地文件,也可以是HDFS
env.registerCachedFile("G:\\aa\\CCC.CSV","localfile");
DataSet<Integer> source=env.fromElements(1,2,3,4,5,6);
source.map(new RichMapFunction<Integer, String>() {
//保存缓存的数据
private String cacheData="";
@Override
public void open(Configuration parameters) throws Exception {
File file = getRuntimeContext().getDistributedCache().getFile("localfile");
List<String> lines = FileUtils.readLines(file);
cacheData=lines.get(0);
}
public String map(Integer value) throws Exception {
return value+cacheData;
}
}).print();
//环境执行
env.execute();
}
}
2、广播变量和分布式缓存道理差不多,它是把一个变量广播出去,而不是文件。使用到的task共用用一份变量,这份变量被保存在每个节点的内存中。这些task可以在一个节点上也可以在不同的节点上。广播变量为了保证各个节点上的数据是一致的只能使用不能修改。
代码:
import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.common.functions.RichMapFunction;
import org.apache.flink.api.java.DataSet;
import org.apache.flink.api.java.ExecutionEnvironment;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.configuration.Configuration;
import scala.Int;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
public class BroadCastDemo {
public static void main(String[] args) throws Exception {
//获取执行环境
ExecutionEnvironment env=ExecutionEnvironment.getExecutionEnvironment();
//创建广播数据源
List<Tuple2<String,Integer>> people=new ArrayList<Tuple2<String,Integer>>();
//添加数据源
people.add(new Tuple2<String, Integer>("Tom",23));
people.add(new Tuple2<String, Integer>("Mary",20));
people.add(new Tuple2<String, Integer>("Mike",26));
DataSet<Tuple2<String, Integer>> peopleData=env.fromCollection(people);
//定义一个需要广播的变量,把Tuple2----> HashMap
DataSet<Map<String,Integer>> peopleBroadCast=
peopleData.map(new MapFunction<Tuple2<String, Integer>, Map<String, Integer>>() {
public Map<String, Integer> map(Tuple2<String, Integer> value) throws Exception {
Map<String, Integer> map = new HashMap<String, Integer>();
map.put(value.f0,value.f1);
return map;
}
});
//执行计算:根据人的姓名来获取年龄
DataSet<String> source=env.fromElements("Tom","Mary","Mike");
source.map(new RichMapFunction<String,String>() {
//定义一个HashMap保存所有广播变量的数据
private HashMap<String,Integer> allMap=new HashMap<String, Integer>();
@Override
public void open(Configuration parameters) throws Exception {
//获取广播变量的数据
List<Map<String,Integer>> data=getRuntimeContext().getBroadcastVariable("mybroadcast");
for (Map<String,Integer> d:data
) {
allMap.putAll(d);
}
}
public String map(String name) throws Exception {
//根据人的姓名获取年龄
Integer age=allMap.get(name);
return "姓名:"+name+"\t 年龄:"+age;
}
}).withBroadcastSet(peopleBroadCast,"mybroadcast").withBroadcastSet(peopleBroadCast, "mybroadcast").print();;
env.execute();
}
}
在广播变量和分布式缓存中在获取文件或是变量的时候都是通过open方法。广播变量可以广播多个变量。广播变量和分布式缓存都适用于小文件和小数据量操作。
3、累加器,累加器是为了在各个节点操作加任务的时候保证数据一致。计数器就是累加器的一种。
代码:
import org.apache.flink.api.common.JobExecutionResult;
import org.apache.flink.api.common.accumulators.IntCounter;
import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.common.functions.RichMapFunction;
import org.apache.flink.api.java.DataSet;
import org.apache.flink.api.java.ExecutionEnvironment;
import org.apache.flink.configuration.Configuration;
public class FlinkWithCounter {
public static void main(String[] args) throws Exception {
//获取执行环境
ExecutionEnvironment env=ExecutionEnvironment.getExecutionEnvironment();
//添加数据源
DataSet<String> data=env.fromElements("Tom","Mary","Mike","Jone").setParallelism(2);
DataSet<Integer> result=data.map(new RichMapFunction<String, Integer>() {
//定义一个累加器(计数器):全局只有一份
private IntCounter intCounter=new IntCounter();
//注册一个累加器
@Override
public void open(Configuration parameters) throws Exception {
this.getRuntimeContext().addAccumulator("myCounter",intCounter);
}
@Override
public Integer map(String s) throws Exception {
this.intCounter.add(1);
return 0;
}
}).setParallelism(2);
result.writeAsText("G:\\aa\\a2.txt");
//获取任务返回结果:获取累加器的结果
//注意:只有任务执行完成了,才能获取累加器的结果
JobExecutionResult execute = env.execute();
int number=execute.getAccumulatorResult("myCounter");
System.out.println("累加的结果是:"+number);
}
}