Example of using Flume HDFS sink with a custom SeqFileFormatter.

MyFormatter.java

package com.foo;

import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.sink.hdfs.SeqFileFormatter;
import org.apache.flume.sink.hdfs.SeqFileFormatter.Record;

public class MyFormatter implements SeqFileFormatter {

  @Override
  public Class<?> getKeyClass() { return Long.class; }

  @Override
  public Class<?> getValueClass() { return String.class; }

  public Iterable<Record> format(Event event) {
    Long key = extractTweetId(event); // TODO implement this
    String value = toXml(event); // TODO implement this
    return Collections.singletonList(new Record(key, value));
  }
 
  public static class Builder implements SeqFileFormatter.Builder {
    @Override
    public SeqFileFormatter build(Context context) {
      // TODO: If you want, you can extract configuration parameters
      // from the context and pass them to your formatter.
      return new MyFormatter();
    }
  }
  
}