Data Sources 是什么呢?就字面意思其实就可以知道:数据来源。
Flink 做为一款流式计算框架,它可用来做批处理,也可以用来做流处理,这个 Data Sources 就是数据的来源地。
flink在批/流****处理中常见的source主要有两大类。
预定义Source
基于本地集合的source(Collection-based-source)
基于文件的source(File-based-source)
基于网络套接字(socketTextStream)
自定义Source
预定义Source演示
Collection [测试]--本地集合Source
在flink最常见的创建DataStream方式有四种:
l 使用env.fromElements(),这种方式也支持Tuple,自定义对象等复合形式。
注意:类型要一致,不一致可以用Object接收,但是使用会报错,比如:env.fromElements("haha", 1);
源码注释中有写:
|使用env.fromCollection(),这种方式支持多种Collection的具体类型,如List,Set,Queue
l 使用env.generateSequence()方法创建基于Sequence的DataStream --已经废弃了
l 使用env.fromSequence()方法创建基于开始和结束的DataStream
一般用于学习测试时编造数据时使用
1.env.fromElements(可变参数);
2.env.fromColletion(各种集合);
3.env.fromSequence(开始,结束);
package com.bigdata.source;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
public class _01YuDingYiSource {
public static void main(String[] args) throws Exception {
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
// 各种获取数据的Source
DataStreamSource<String> dataStreamSource = env.fromElements("hello world txt", "hello nihao kongniqiwa");
dataStreamSource.print();
// 演示一个错误的
//DataStreamSource<Object> dataStreamSource2 = env.fromElements("hello", 1,3.0f);
//dataStreamSource2.print();
DataStreamSource<Tuple2<String, Integer>> elements = env.fromElements(
Tuple2.of("张三", 18),
Tuple2.of("lisi", 18),
Tuple2.of("wangwu", 18)
);
elements.print();
// 有一个方法,可以直接将数组变为集合 复习一下数组和集合以及一些非常常见的API
String[] arr = {"hello","world"};
System.out.println(arr.length);
System.out.println(Arrays.toString(arr));
List<String> list = Arrays.asList(arr);
System.out.println(list);
env.fromElements(
Arrays.asList(arr),
Arrays.asList(arr),
Arrays.asList(arr)
).print();
// 第二种加载数据的方式
// Collection 的子接口只有 Set 和 List
ArrayList<String> list1 = new ArrayList<>();
list1.add("python");
list1.add("scala");
list1.add("java");
DataStreamSource<String> ds1 = env.fromCollection(list1);
DataStreamSource<String> ds2 = env.fromCollection(Arrays.asList(arr));
// 第三种
DataStreamSource<Long> ds3 = env.fromSequence(1, 100);
ds3.print();
// execute 下面的代码不运行,所以,这句话要放在最后。
env.execute("获取预定义的Source");
}
}
本地文件的案例:
package com.bigdata.source;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import java.io.File;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
public class _02YuDingYiSource {
public static void main(String[] args) throws Exception {
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
// 获取并行度
System.out.println(env.getParallelism());
// 讲第二种Source File类型的
// 给了一个相对路径,说路径不对,老闫非要写,我咋办?
// 相对路径,转绝对路径
File file = new File("datas/wc.txt");
File file2 = new File("./");
System.out.println(file.getAbsoluteFile());
System.out.println(file2.getAbsoluteFile());
DataStreamSource<String> ds1 = env.readTextFile("datas/wc.txt");
ds1.print();
// 还可以获取hdfs路径上的数据
DataStreamSource<String> ds2 = env.readTextFile("hdfs://bigdata01:9820/home/a.txt");
ds2.print();
// execute 下面的代码不运行,所以,这句话要放在最后。
env.execute("获取预定义的Source");
}
}
Socket [测试]
socketTextStream(String hostname, int port) 方法是一个非并行的Source,该方法需要传入两个参数,第一个是指定的IP地址或主机名,第二个是端口号,即从指定的Socket读取数据创建DataStream。该方法还有多个重载的方法,其中一个是socketTextStream(String hostname, int port, String delimiter, long maxRetry),这个重载的方法可以指定行分隔符和最大重新连接次数。这两个参数,默认行分隔符是”\n”,最大重新连接次数为0。
提示:
如果使用socketTextStream读取数据,在启动Flink程序之前,必须先启动一个Socket服务,为了方便,Mac或Linux用户可以在命令行终端输入nc -lk 8888启动一个Socket服务并在命令行中向该Socket服务发送数据。Windows用户可以在百度中搜索windows安装netcat命令。
使用nc 进行数据的发送
yum install -y nc
nc -lk 8888 --向8888端口发送消息,这个命令先运行,如果先运行java程序,会报错!
如果是windows平台:nc -lp 8888
代码演示:
//socketTextStream创建的DataStream,不论怎样,并行度永远是1
public class StreamSocketSource {
public static void main(String[] args) throws Exception {
//local模式默认的并行度是当前机器的逻辑核的数量
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
int parallelism0 = env.getParallelism();
System.out.println("执行环境默认的并行度:" + parallelism0);
DataStreamSource<String> lines = env.socketTextStream("localhost", 8888);
//获取DataStream的并行度
int parallelism = lines.getParallelism();
System.out.println("SocketSource的并行度:" + parallelism);
SingleOutputStreamOperator<String> words = lines.flatMap(new FlatMapFunction<String, String>() {
@Override
public void flatMap(String line, Collector<String> collector) throws Exception {
String[] words = line.split(" ");
for (String word : words) {
collector.collect(word);
}
}
});
int parallelism2 = words.getParallelism();
System.out.println("调用完FlatMap后DataStream的并行度:" + parallelism2);
words.print();
env.execute();
}
}
以下用于演示:统计socket中的 单词数量,体会流式计算的魅力!
import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.api.common.functions.FlatMapFunction;
import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.java.functions.KeySelector;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.datastream.KeyedStream;
import org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.util.Collector;
public class SourceDemo02_Socket {
public static void main(String[] args) throws Exception {
//TODO 1.env-准备环境
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);
//TODO 2.source-加载数据
DataStream<String> socketDS = env.socketTextStream("bigdata01", 8889);
//TODO 3.transformation-数据转换处理
//3.1对每一行数据进行分割并压扁
DataStream<String> wordsDS = socketDS.flatMap(new FlatMapFunction<String, String>() {
@Override
public void flatMap(String value, Collector<String> out) throws Exception {
String[] words = value.split(" ");
for (String word : words) {
out.collect(word);
}
}
});
//3.2每个单词记为<单词,1>
DataStream<Tuple2<String, Integer>> wordAndOneDS = wordsDS.map(new MapFunction<String, Tuple2<String, Integer>>() {
@Override
public Tuple2<String, Integer> map(String value) throws Exception {
return Tuple2.of(value, 1);
}
});
//3.3分组
KeyedStream<Tuple2<String, Integer>, String> keyedDS = wordAndOneDS.keyBy(new KeySelector<Tuple2<String, Integer>, String>() {
@Override
public String getKey(Tuple2<String, Integer> value) throws Exception {
return value.f0;
}
});
//3.4聚合
SingleOutputStreamOperator<Tuple2<String, Integer>> result = keyedDS.sum(1);
//TODO 4.sink-数据输出
result.print();
//TODO 5.execute-执行
env.execute();
}
}
自定义数据源
SourceFunction:非并行数据源(并行度只能=1) --接口
RichSourceFunction:多功能非并行数据源(并行度只能=1) --类
ParallelSourceFunction:并行数据源(并行度能够>=1) --接口
RichParallelSourceFunction:多功能并行数据源(并行度能够>=1) --类 【建议使用的】
package com.bigdata.day02;
import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.NoArgsConstructor;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.source.ParallelSourceFunction;
import org.apache.flink.streaming.api.functions.source.RichParallelSourceFunction;
import org.apache.flink.streaming.api.functions.source.RichSourceFunction;
import org.apache.flink.streaming.api.functions.source.SourceFunction;
import java.util.Random;
import java.util.UUID;
/**
* 需求: 每隔1秒随机生成一条订单信息(订单ID、用户ID、订单金额、时间戳)
* 要求:
* - 随机生成订单ID(UUID)
* - 随机生成用户ID(0-2)
* - 随机生成订单金额(0-100)
* - 时间戳为当前系统时间
*/
@Data // set get toString
@AllArgsConstructor
@NoArgsConstructor
class OrderInfo{
private String orderId;
private int uid;
private int money;
private long timeStamp;
}
// class MySource extends RichSourceFunction<OrderInfo> {
//class MySource extends RichParallelSourceFunction<OrderInfo> {
class MySource implements SourceFunction<OrderInfo> {
boolean flag = true;
@Override
public void run(SourceContext ctx) throws Exception {
// 源源不断的产生数据
Random random = new Random();
while(flag){
OrderInfo orderInfo = new OrderInfo();
orderInfo.setOrderId(UUID.randomUUID().toString());
orderInfo.setUid(random.nextInt(3));
orderInfo.setMoney(random.nextInt(101));
orderInfo.setTimeStamp(System.currentTimeMillis());
ctx.collect(orderInfo);
Thread.sleep(1000);// 间隔1s
}
}
// source 停止之前需要干点啥
@Override
public void cancel() {
flag = false;
}
}
public class CustomSource {
public static void main(String[] args) throws Exception {
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
env.setParallelism(2);
// 将自定义的数据源放入到env中
DataStreamSource dataStreamSource = env.addSource(new MySource())/*.setParallelism(1)*/;
System.out.println(dataStreamSource.getParallelism());
dataStreamSource.print();
env.execute();
}
}
通过****ParallelSourceFunction创建可并行Source
/**
* 自定义多并行度Source
*/
public class CustomerSourceWithParallelDemo {
public static void main(String[] args) throws Exception {
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
DataStreamSource<String> mySource = env.addSource(new MySource()).setParallelism(6);
mySource.print();
env.execute();
}
public static class MySource implements ParallelSourceFunction<String> {
@Override
public void run(SourceContext<String> ctx) throws Exception {
ctx.collect(UUID.randomUUID().toString());
/*
如果不设置无限循环可以看出,设置了多少并行度就打印出多少条数据
*/
}
@Override
public void cancel() {}
}
}
如果代码换成ParallelSourceFunction,每次生成12个数据,假如是12核数的话。
总结:Rich富函数总结 ctrl + o
Rich 类型的Source可以比非Rich的多出有:
- open方法,实例化的时候会执行一次,多个并行度会执行多次的哦(因为是多个实例了)
- close方法,销毁实例的时候会执行一次,多个并行度会执行多次的哦
- getRuntimeContext 方法可以获得当前的Runtime对象(底层API)
Kafka Source --从kafka中读取数据
https://nightlies.apache.org/flink/flink-docs-release-1.13/zh/docs/connectors/datastream/kafka/
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-connector-kafka_2.11</artifactId>
<version>${flink.version}</version>
</dependency>
创建一个topic1 这个主题:
cd /opt/installs/kafka3/
bin/kafka-topics.sh --bootstrap-server bigdata01:9092 --create --partitions 1 --replication-factor 3 --topic topic1
通过控制台向topic1发送消息:
bin/kafka-console-producer.sh --bootstrap-server bigdata01:9092 --topic topic1
package com.bigdata.day02;
import org.apache.flink.api.common.functions.FilterFunction;
import org.apache.flink.api.common.serialization.SimpleStringSchema;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer;
import java.util.Properties;
public class KafkaSource {
public static void main(String[] args) throws Exception {
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
Properties properties = new Properties();
properties.setProperty("bootstrap.servers", "bigdata01:9092");
properties.setProperty("group.id", "g1");
FlinkKafkaConsumer<String> kafkaSource = new FlinkKafkaConsumer<String>("topic1",new SimpleStringSchema(),properties);
DataStreamSource<String> dataStreamSource = env.addSource(kafkaSource);
// 以下代码跟flink消费kakfa数据没关系,仅仅是将需求搞的复杂一点而已
// 返回true 的数据就保留下来,返回false 直接丢弃
dataStreamSource.filter(new FilterFunction<String>() {
@Override
public boolean filter(String word) throws Exception {
// 查看单词中是否包含success 字样
return word.contains("success");
}
}).print();
env.execute();
}
}
版权归原作者 时差953 所有, 如有侵权,请联系我们删除。