先说说partition的好处：Partition的好处是可以并发的获取同类数据，提高效率。
第一步需要实现Partitioner对象。
public class ProducerPartitioner implements Partitioner<String> { public static final Logger LOG=LoggerFactory.getLogger(UserInfo.class); @Override public int partition(String key, int numPartitions) { LOG.info("ProducerPartitioner key:"+key+" partitions:"+numPartitions); return key.length() % numPartitions; }}/**选择用哪个类来进行设置partition*/props.put("partitioner.class", "org.kafka.partitioner.ProducerPartitioner");ProducerData<String,UserInfo> data=new ProducerData<String, UserInfo>("test","developerTest", msg);原文：http://www.cnblogs.com/xubiao/p/5361773.html