爲什麼將CSV的數據發到kafka
- flink做流式計算時,選用kafka消息作爲數據源是常用手段,因此在學習和開發flink過程中,也會將數據集文件中的記錄發送到kafka,來模擬不間斷數據;
- 整個流程如下:
- 您可能會覺得這樣做多此一舉:flink直接讀取CSV不就行了嗎?這樣做的原因如下:
- 首先,這是學習和開發時的做法,數據集是CSV文件,而生產環境的實時數據卻是kafka數據源;
- 其次,Java應用中可以加入一些特殊邏輯,例如數據處理,彙總統計(用來和flink結果對比驗證);
- 另外,如果兩條記錄實際的間隔時間如果是1分鐘,那麼Java應用在發送消息時也可以間隔一分鐘再發送,這個邏輯在flink社區的demo中有具體的實現,此demo也是將數據集發送到kafka,再由flink消費kafka,地址是:https://github.com/ververica/sql-training
如何將CSV的數據發送到kafka
前面的圖可以看出,讀取CSV再發送消息到kafka的操作是Java應用所爲,因此今天的主要工作就是開發這個Java應用,並驗證;
版本信息
- JDK:1.8.0_181
- 開發工具:IntelliJ IDEA 2019.2.1 (Ultimate Edition)
- 開發環境:Win10
- Zookeeper:3.4.13
- Kafka:2.4.0(scala:2.12)
關於數據集
- 本次實戰用到的數據集是CSV文件,裏面是一百零四萬條淘寶用戶行爲數據,該數據來源是阿里雲天池公開數據集,我對此數據做了少量調整;
- 此CSV文件可以在CSDN下載,地址:https://download.csdn.net/download/boling_cavalry/12381698
- 也可以在我的Github下載,地址:https://raw.githubusercontent.com/zq2599/blog_demos/master/files/UserBehavior.7z
- 該CSV文件的內容,一共有六列,每列的含義如下表:
列名稱 | 說明 |
---|---|
用戶ID | 整數類型,序列化後的用戶ID |
商品ID | 整數類型,序列化後的商品ID |
商品類目ID | 整數類型,序列化後的商品所屬類目ID |
行爲類型 | 字符串,枚舉類型,包括(‘pv’, ‘buy’, ‘cart’, ‘fav’) |
時間戳 | 行爲發生的時間戳 |
時間字符串 | 根據時間戳字段生成的時間字符串 |
- 關於該數據集的詳情,請參考《準備數據集用於flink學習》
Java應用簡介
編碼前,先把具體內容列出來,然後再挨個實現:
- 從CSV讀取記錄的工具類:UserBehaviorCsvFileReader
- 每條記錄對應的Bean類:UserBehavior
- Java對象序列化成JSON的序列化類:JsonSerializer
- 向kafka發送消息的工具類:KafkaProducer
- 應用類,程序入口:SendMessageApplication
上述五個類即可完成Java應用的工作,接下來開始編碼吧;
直接下載源碼
- 如果您不想寫代碼,您可以直接從GitHub下載這個工程的源碼,地址和鏈接信息如下表所示:
名稱 | 鏈接 | 備註 |
---|---|---|
項目主頁 | https://github.com/zq2599/blog_demos | 該項目在GitHub上的主頁 |
git倉庫地址(https) | https://github.com/zq2599/blog_demos.git | 該項目源碼的倉庫地址,https協議 |
git倉庫地址(ssh) | [email protected]:zq2599/blog_demos.git | 該項目源碼的倉庫地址,ssh協議 |
- 這個git項目中有多個文件夾,本章源碼在flinksql這個文件夾下,如下圖紅框所示:
編碼
- 創建maven工程,pom.xml如下,比較重要的jackson和javacsv的依賴:
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0</modelVersion>
<groupId>com.bolingcavalry</groupId>
<artifactId>flinksql</artifactId>
<version>1.0-SNAPSHOT</version>
<properties>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
<flink.version>1.10.0</flink.version>
<kafka.version>2.2.0</kafka.version>
<java.version>1.8</java.version>
<scala.binary.version>2.11</scala.binary.version>
<maven.compiler.source>${java.version}</maven.compiler.source>
<maven.compiler.target>${java.version}</maven.compiler.target>
</properties>
<dependencies>
<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka-clients</artifactId>
<version>${kafka.version}</version>
</dependency>
<dependency>
<groupId>com.fasterxml.jackson.core</groupId>
<artifactId>jackson-databind</artifactId>
<version>2.9.10.1</version>
</dependency>
<!-- Logging dependencies -->
<dependency>
<groupId>org.slf4j</groupId>
<artifactId>slf4j-log4j12</artifactId>
<version>1.7.7</version>
<scope>runtime</scope>
</dependency>
<dependency>
<groupId>log4j</groupId>
<artifactId>log4j</artifactId>
<version>1.2.17</version>
<scope>runtime</scope>
</dependency>
<dependency>
<groupId>net.sourceforge.javacsv</groupId>
<artifactId>javacsv</artifactId>
<version>2.0</version>
</dependency>
</dependencies>
<build>
<plugins>
<!-- Java Compiler -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-compiler-plugin</artifactId>
<version>3.1</version>
<configuration>
<source>${java.version}</source>
<target>${java.version}</target>
</configuration>
</plugin>
<!-- Shade plugin to include all dependencies -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-shade-plugin</artifactId>
<version>3.0.0</version>
<executions>
<!-- Run shade goal on package phase -->
<execution>
<phase>package</phase>
<goals>
<goal>shade</goal>
</goals>
<configuration>
<artifactSet>
<excludes>
</excludes>
</artifactSet>
<filters>
<filter>
<!-- Do not copy the signatures in the META-INF folder.
Otherwise, this might cause SecurityExceptions when using the JAR. -->
<artifact>*:*</artifact>
<excludes>
<exclude>META-INF/*.SF</exclude>
<exclude>META-INF/*.DSA</exclude>
<exclude>META-INF/*.RSA</exclude>
</excludes>
</filter>
</filters>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</build>
</project>
- 從CSV讀取記錄的工具類:UserBehaviorCsvFileReader,後面在主程序中會用到java8的Steam API來處理集合,所以UserBehaviorCsvFileReader實現了Supplier接口:
public class UserBehaviorCsvFileReader implements Supplier<UserBehavior> {
private final String filePath;
private CsvReader csvReader;
public UserBehaviorCsvFileReader(String filePath) throws IOException {
this.filePath = filePath;
try {
csvReader = new CsvReader(filePath);
csvReader.readHeaders();
} catch (IOException e) {
throw new IOException("Error reading TaxiRecords from file: " + filePath, e);
}
}
@Override
public UserBehavior get() {
UserBehavior userBehavior = null;
try{
if(csvReader.readRecord()) {
csvReader.getRawRecord();
userBehavior = new UserBehavior(
Long.valueOf(csvReader.get(0)),
Long.valueOf(csvReader.get(1)),
Long.valueOf(csvReader.get(2)),
csvReader.get(3),
new Date(Long.valueOf(csvReader.get(4))*1000L));
}
} catch (IOException e) {
throw new NoSuchElementException("IOException from " + filePath);
}
if (null==userBehavior) {
throw new NoSuchElementException("All records read from " + filePath);
}
return userBehavior;
}
}
- 每條記錄對應的Bean類:UserBehavior,和CSV記錄格式保持一致即可,表示時間的ts字段,使用了JsonFormat註解,在序列化的時候以此來控制格式:
public class UserBehavior {
@JsonFormat
private long user_id;
@JsonFormat
private long item_id;
@JsonFormat
private long category_id;
@JsonFormat
private String behavior;
@JsonFormat(shape = JsonFormat.Shape.STRING, pattern = "yyyy-MM-dd'T'HH:mm:ss'Z'")
private Date ts;
public UserBehavior() {
}
public UserBehavior(long user_id, long item_id, long category_id, String behavior, Date ts) {
this.user_id = user_id;
this.item_id = item_id;
this.category_id = category_id;
this.behavior = behavior;
this.ts = ts;
}
}
- Java對象序列化成JSON的序列化類:JsonSerializer
public class JsonSerializer<T> {
private final ObjectMapper jsonMapper = new ObjectMapper();
public String toJSONString(T r) {
try {
return jsonMapper.writeValueAsString(r);
} catch (JsonProcessingException e) {
throw new IllegalArgumentException("Could not serialize record: " + r, e);
}
}
public byte[] toJSONBytes(T r) {
try {
return jsonMapper.writeValueAsBytes(r);
} catch (JsonProcessingException e) {
throw new IllegalArgumentException("Could not serialize record: " + r, e);
}
}
}
- 向kafka發送消息的工具類:KafkaProducer:
public class KafkaProducer implements Consumer<UserBehavior> {
private final String topic;
private final org.apache.kafka.clients.producer.KafkaProducer<byte[], byte[]> producer;
private final JsonSerializer<UserBehavior> serializer;
public KafkaProducer(String kafkaTopic, String kafkaBrokers) {
this.topic = kafkaTopic;
this.producer = new org.apache.kafka.clients.producer.KafkaProducer<>(createKafkaProperties(kafkaBrokers));
this.serializer = new JsonSerializer<>();
}
@Override
public void accept(UserBehavior record) {
// 將對象序列化成byte數組
byte[] data = serializer.toJSONBytes(record);
// 封裝
ProducerRecord<byte[], byte[]> kafkaRecord = new ProducerRecord<>(topic, data);
// 發送
producer.send(kafkaRecord);
// 通過sleep控制消息的速度,請依據自身kafka配置以及flink服務器配置來調整
try {
Thread.sleep(500);
}catch(InterruptedException e){
e.printStackTrace();
}
}
/**
* kafka配置
* @param brokers The brokers to connect to.
* @return A Kafka producer configuration.
*/
private static Properties createKafkaProperties(String brokers) {
Properties kafkaProps = new Properties();
kafkaProps.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, brokers);
kafkaProps.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, ByteArraySerializer.class.getCanonicalName());
kafkaProps.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, ByteArraySerializer.class.getCanonicalName());
return kafkaProps;
}
}
- 最後是應用類SendMessageApplication,CSV文件路徑、kafka的topic和borker地址都在此設置,另外借助java8的Stream API,只需少量代碼即可完成所有工作:
public class SendMessageApplication {
public static void main(String[] args) throws Exception {
// 文件地址
String filePath = "D:\\temp\\202005\\02\\UserBehavior.csv";
// kafka topic
String topic = "user_behavior";
// kafka borker地址
String broker = "192.168.50.43:9092";
Stream.generate(new UserBehaviorCsvFileReader(filePath))
.sequential()
.forEachOrdered(new KafkaProducer(topic, broker));
}
}
驗證
- 請確保kafka已經就緒,並且名爲user_behavior的topic已經創建;
- 請將CSV文件準備好;
- 確認SendMessageApplication.java中的文件地址、kafka topic、kafka broker三個參數準確無誤;
- 運行SendMessageApplication.java;
- 開啓一個 控制檯消息kafka消息,參考命令如下:
./kafka-console-consumer.sh \
--bootstrap-server 127.0.0.1:9092 \
--topic user_behavior \
--consumer-property group.id=old-consumer-test \
--consumer-property consumer.id=old-consumer-cl \
--from-beginning
- 正常情況下可以立即見到消息,如下圖:
至此,通過Java應用模擬用戶行爲消息流的操作就完成了,接下來的flink實戰就用這個作爲數據源;