Question

我想在foreachRDD方法中检索RDD的每条记录上的每个kafka偏移量。我的主题中有一个分区，所以我的RDD也得到一个分区。我基本上尝试过这样的事情：

dStream.foreachRDD { rdd =>
  if (!rdd.isEmpty) {
    //get offset first value of the offset
    val firstOffset = rdd.asInstanceOf[HasOffsetRanges].offsetRanges(0).fromOffset
    val rddWithOffset = rdd.map(_.value)
      .zipWithIndex()
      .map{ case (v,i) => (v,i + firstOffset)}
  }
}

例如，在我的生产者中，我使用循环发送消息，并将索引放在名为position的列中，如下所示：

+------+-----+--------+
|  name|  age|position|
+------+-----+--------+
|johnny|   26|       1|
| chloe|   42|       2|
| brian|   19|       3|
| eliot|   35|       4|
+------+-----+--------+

不幸地，我注意到当我在消费者中添加偏移列时，订单不会被维护：

+------+-----+--------+------+
|  name|  age|position|offset|
+------+-----+--------+------+
|johnny|   26|       1|     1|
| chloe|   42|       2|     3|
| brian|   19|       3|     4|
| eliot|   35|       4|     2|
+------+-----+--------+------+

似乎我通过这个过程放松了订单。你有什么主意吗？感谢

顺便说一下，我的Java制作人看起来像这样：

KafkaRestProducer<String, Object> producer = new KafkaRestProducer<>(props);

ArrayList<String> names = new ArrayList<String>()
names.add("johnny")
names.add("chloe")
names.add("brian")
names.add("eliot")

ArrayList<Integer> ages = ArrayList<Integer>()
names.add(26)
names.add(42)
names.add(19)
names.add(35)

for (int i = 0; i < 3; ++i) {

    String name = names(i)
    Int age = ages(i)     
    Person person = Person
        .newBuilder()
        .setName(name)
        .setAge(age)
        .setPosition(i)
        .build();

    ProducerRecord<String, Object> record = new ProducerRecord<>("/apps/PERSON/streams:myTopic", name, person);

    producer.send(record, null);
    System.out.println(i);
}

Answer 1

我的英语很差。我使用这段代码：

    val Array(brokers, topic, groupId) = args
    val kafkaParams = Map[String, String]("metadata.broker.list" -> brokers, "group.id" -> groupId)
    val topicPartition = Map[TopicAndPartition, Long](TopicAndPartition(topic, 0) -> 1.toLong)
    val messageHandler = (mmd: MessageAndMetadata[String, String]) => (mmd.offset, mmd.message)
    val kafkaStream = KafkaUtils.createDirectStream[String, String, StringDecoder, StringDecoder, (Long, String)](
        ssc, kafkaParams, topicPartition, messageHandler)

    kafkaStream.foreachRDD(rdd => rdd.foreach(println))

输出：（offset，lineOfMessage） ...

将kafka偏移量附加到foreachRDD中的每条记录

1 个答案: