Java代码中Kafka的使用

kafka的客户端，建立消费者

public class KafkaClient {    public static ConsumerConnector createConsumer() {
        Properties consumerProducer = new Properties();
        consumerProducer.put("zookeeper.connect", "ip:port");
        consumerProducer.put("group.id", "group-1");
        consumerProducer.put("serializer.class", "kafka.serializer.StringEncoder");
        consumerProducer .put("auto.offset.reset", "smallest");
        ConsumerConfig config = new ConsumerConfig(consumerProducer);
        ConsumerConnector consumer = Consumer.createJavaConsumerConnector(config);        return consumer;
    }    public static class ConsumerTest extends Thread {
        AtomicBoolean start = new AtomicBoolean(true);
        CountDownLatch c;        private String topic;
        ConsumerConnector consumer;        public ConsumerTest(String topic, ConsumerConnector consumer) {            this.consumer = consumer;            this.topic = topic;
        }        public void run() {
            Map<String, Integer> topicCountMap = new HashMap<String, Integer>();
            topicCountMap.put(topic, 1);

            Map<String, List<KafkaStream<byte[], byte[]>>> createMessageStreams = consumer.createMessageStreams(topicCountMap);
            KafkaStream<byte[], byte[]> stream = createMessageStreams.get(topic).get(0);// 获取每次接收到的这个数据

            ConsumerIterator<byte[], byte[]> iterator = stream.iterator();            while (iterator.hasNext()) {
                String j = new String(iterator.next().message());
                System.out.println(j);
            }

        }

    }    public static void main(String[] args) throws InterruptedException {
        ConsumerConnector c = KafkaClient.createConsumer();
        ConsumerTest t = new ConsumerTest("topic_sname", c);
        t.start();
        Thread.sleep(1000 * 60 * 10);
        t.interrupt();
        c.shutdown();
    }
}

kafka数据收集

public class KafkaDataCollect {    static Logger LOG = Logger.getLogger(KafkaDataCollect.class);
    ConsumerConnector consumer = KafkaClient.createConsumer();    //尺度时间戳 用于限定每日更新量
    private long ruler = 0;
    KafkaDataCollect(){        //初始化尺度时间戳
        ruler = getRuler();
    }    public long getRuler(){        long current = System.currentTimeMillis();
        String date = new java.text.SimpleDateFormat("dd/MM/yyyy").format(new java.util.Date(current));
        date = date + " 00:00:00";        long rulertime = 0; 
        try {
            rulertime = new java.text.SimpleDateFormat("dd/MM/yyyy HH:mm:ss").parse(date).getTime();
        } catch (ParseException e) {            // TODO Auto-generated catch block
            e.printStackTrace();
        }        return rulertime;
    }    
    public void dataStreamIn(String topic) {
        Map<String, Integer> topicCountMap = new HashMap<String, Integer>();
        topicCountMap.put(topic, 1);

        Map<String, List<KafkaStream<byte[], byte[]>>> createMessageStreams = consumer.createMessageStreams(topicCountMap);
        KafkaStream<byte[], byte[]> stream = createMessageStreams.get(topic).get(0);// 获取每次接收到的这个数据
        int articleCount = 0;
        ConsumerIterator<byte[], byte[]> iterator = stream.iterator();
        Jedis jedis = new Jedis(ip, 6379);        try {            while (iterator.hasNext()) {
                String msgStr = "";                if((System.currentTimeMillis() - ruler) >= 24 * 60 * 60 * 1000)
                {
                    String date = new java.text.SimpleDateFormat("dd/MM/yyyy").format(new java.util.Date(ruler));
                    msgStr = date + "Kafka data stream collected "+ articleCount+ " articles.";
                    LOG.info(msgStr);
                    articleCount = 0;
                    ruler = ruler + 24 * 60 * 60 * 1000;
                }                if((System.currentTimeMillis() - ruler) >= 8 * 60 * 60 * 1000)
                {
                    SendMessage.send("17865153777", msgStr, null, null, null);
                    msgStr = "";
                }
                String j = new String(iterator.next().message());
                Gson gson = new Gson();
                JsonFromKafkaData jsonD = gson.fromJson(j, new TypeToken<JsonFromKafkaData>() {
                }.getType());                try {
                    LOG.info(j);                    if(jsonD.getSimId()!=null && !jsonD.getSimId().contains("_"))
                        jsonD.setSimId("clusterId_"+jsonD.getSimId());
                    jedis.lpush("kafka-queue", gson.toJson(jsonD));
                    articleCount++;
                } catch (Exception e) {                    // TODO: handle exception
                    LOG.error("Input newData into queue.", e);
                }
            }
        } catch (Exception e) {
            jedis.close();
            SendMessage.send("178xxxxxxxx", "Kafka data collection process stoped.", null, null, null);            // TODO: handle exception
            e.printStackTrace();
        }
    }    
    
    public static void main(String[] args) {
        KafkaDataCollect kafkaStream = new KafkaDataCollect();
        kafkaStream.dataStreamIn("topic_name");
    }

}

作者：牛粪
链接：https://www.jianshu.com/p/a564932d91aa