嗨伙计们，我有下一个问题。我正在使用带有Java的Apache Spark Streaming v1.6.0从IBM MQ获得一些消息。我为MQ制作了自定义接收器，但我遇到的问题是我需要将RDD从JavaDStream转换为DataFrame。为此，我使用foreachRDD迭代JavaDStream并且我定义了DataFrame的模式，但是当我运行作业时，第一个消息会抛出下一个异常：
java.lang.ClassCastException：org.apache.spark.rdd.BlockRDDPartition无法在org.apache.spark.rdd.ParallelCollectionRDD.compute（ParallelCollectionRDD.scala：102）中强制转换为org.apache.spark.rdd.ParallelCollectionPartition。 aplet.spark.rdd.RDD.computeOrReadCheckpoint（RDD.scala：306）atg.apache.spark.rdd.RDD.iterator（RDD.scala：270）at org.apache.spark.scheduler.ResultTask.runTask（ResultTask。 scala：66）atg.apache.spark.scheduler.Task.run（Task.scala：89）at org.apache.spark.executor.Executor $ TaskRunner.run（Executor.scala：213）at java.util.concurrent .ThreadPoolExecutor.runWorker（ThreadPoolExecutor.java:1149）java.util.concurrent.ThreadPoolExecutor $ Worker.run（ThreadPoolExecutor.java:624）at java.lang.Thread.run（Thread.java:748)19/03/28 12:53:26 WARN TaskSetManager：阶段0.0中失去的任务0.0（TID 0，localhost）：java.lang.ClassCastException：org.apache.spark.rdd.BlockRDDPartition无法强制转换为org.apache.spark.rdd.ParallelCollectionPartition at org.apache.spark.rdd.Parallel收集RDD.compute（ParallelCollectionRDD.scala：102）位于org.apache.spark.rdd.RDD.computeOrReadCheckpoint（RDD.scala：306）atg.apache.spark.rdd.RDD.iterator（RDD.scala：270）org .apache.spark.scheduler.ResultTask.runTask（ResultTask.scala：66）atg.apache.spark.scheduler.Task.run（Task.scala：89）at org.apache.spark.executor.Executor $ TaskRunner.run （Executor.scala：213）位于java.util.concurrent.ThreadPoolExecutor.runWorker（ThreadPoolExecutor.java:1149）的java.util.concurrent.ThreadPoolExecutor $ Worker.run（ThreadPoolExecutor.java:624），位于java.lang.Thread。运行（Thread.java:748）
然后代码执行得很好。即使我在MQ中没有任何消息，也只是当我运行de job时的第一条消息。
这是我的CustomMQReceiver
public CustomMQReceiver() {
super(StorageLevel.MEMORY_ONLY_2());
}
@Override
public void onStart() {
new Thread() {
@Override
public void run() {
try {
initConnection();
receive();
} catch (JMSException ex) {
ex.printStackTrace();
}
}
}.start();
}
@Override
public void onStop() {
}
private void receive() {
System.out.print("Started receiving messages from MQ");
try {
Message receivedMessage = null;
while (!isStopped() && (receivedMessage = consumer.receiveNoWait()) != null) {
String userInput = convertStreamToString(receivedMessage);
System.out.println("Received data :'" + userInput + "'");
store(userInput);
}
stop("No More Messages To read !");
qCon.close();
System.out.println("Queue Connection is Closed");
} catch (Exception e) {
e.printStackTrace();
restart("Trying to connect again");
} catch (Throwable t) {
restart("Error receiving data", t);
}
}
public void initConnection() throws JMSException {
MQQueueConnectionFactory conFactory = new MQQueueConnectionFactory();
conFactory.setHostName(HOST);
conFactory.setPort(PORT);
conFactory.setIntProperty(WMQConstants.WMQ_CONNECTION_MODE, WMQConstants.WMQ_CM_CLIENT);
conFactory.setQueueManager(QMGR);
conFactory.setChannel(CHANNEL);
conFactory.setBooleanProperty(WMQConstants.USER_AUTHENTICATION_MQCSP, true);
conFactory.setStringProperty(WMQConstants.USERID, APP_USER);
conFactory.setStringProperty(WMQConstants.PASSWORD, APP_PASSWORD);
qCon = (MQQueueConnection) conFactory.createConnection();
MQQueueSession qSession = (MQQueueSession) qCon.createQueueSession(false, 1);
MQQueue queue = (MQQueue) qSession.createQueue(QUEUE_NAME);
consumer = (MQMessageConsumer) qSession.createConsumer(queue);
qCon.start();
}
@Override
public StorageLevel storageLevel() {
return StorageLevel.MEMORY_ONLY_2();
}
private static String convertStreamToString(final Message jmsMsg) throws Exception {
String stringMessage = "";
JMSTextMessage msg = (JMSTextMessage) jmsMsg;
stringMessage = msg.getText();
return stringMessage;
}
这是我的火花代码
SparkConf sparkConf = new SparkConf()
.setAppName("MQStreaming")
.set("spark.driver.allowMultipleContexts", "true")
.setMaster("local[*]");
JavaSparkContext jsc = new JavaSparkContext(sparkConf);
final SQLContext sqlContext = new SQLContext(jsc);
JavaStreamingContext ssc = new JavaStreamingContext(sparkConf, new Duration(Long.parseLong(propertiesConf.getProperty("duration"))));
JavaDStream<String> customReceiverStream = ssc.receiverStream(new CustomMQReceiver());
customReceiverStream.foreachRDD(new VoidFunction<JavaRDD<String>>() {
@Override
public void call(JavaRDD<String> rdd) throws Exception {
JavaRDD<Row> rddRow = rdd.map(new Function<String, Row>() {
@Override
public Row call(String v1) throws Exception {
return RowFactory.create(v1);
}
});
try {
StructType schema = new StructType(new StructField[]{
new StructField("trama", DataTypes.StringType, true, Metadata.empty())
});
DataFrame frame = sqlContext.createDataFrame(rddRow, schema);
if (frame.count() > 0) {
//Here is where the first messages throw the exception
frame.show();
frame.write().mode(SaveMode.Append).json("file:///C:/tmp/");
}
} catch (Exception ex) {
System.out.println(" INFO " + ex.getMessage());
}
}
});
ssc.start();
ssc.awaitTermination();
我无法更改spark的版本，因为这个作业将在一个带有spark 1.6的旧cloudera集群中运行。我不知道我做错了什么或者只是一个错误。救命！！！！
我解决了自己的问题，这个异常是由我如何创建SQLContext抛出的，正确的方法是使用JavaStreamingContext创建sqlContext
//JavaStreamingContext jsc = ...
SQLContext sqlContext = new SQLContext(jsc.sparkContext());