HBaseFill.java

import java.io.*;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.*;
import org.apache.hadoop.hbase.client.*;
import org.apache.hadoop.hbase.util.*;

public class HBaseFill {
  private static void createTable(final HBaseAdmin admin, final byte[] tableName, 
      final byte[]... families) throws IOException {
    HTableDescriptor htd = new HTableDescriptor(tableName);
    for (byte[] family: families) {
      HColumnDescriptor hcd = new HColumnDescriptor(family);
      htd.addFamily(hcd);
    }
    byte[][] splitKeys = new byte[16][];
    byte[] hex = Bytes.toBytes("0123456789abcdef");
    for (int i = 0; i < 16; ++i) {
      splitKeys[i] = new byte[] { hex[i] };
    }
    admin.createTable(htd, splitKeys);
  }

  public static void loadData(final HTable table, int rows, byte[]... families) 
      throws IOException {
    byte[] qualifier = Bytes.toBytes("q");
    table.setAutoFlush(false);
    for (int i = 0; i < rows; ++i) {
      byte[] value = Bytes.add(Bytes.toBytes(System.currentTimeMillis()), Bytes.toBytes(rows));
      byte[] key = Bytes.toBytes(MD5Hash.getMD5AsHex(value));
      Put put = new Put(key);
      put.setWriteToWAL(false);
      for (byte[] family: families) {
        put.add(family, qualifier, value);
      }
      table.put(put);

      if ((i % 5000) == 0) {
        System.out.println("rows inserted " + i);
        table.flushCommits();
      }
    }
    table.flushCommits();
  }

  public static void main(String[] args) throws Exception {
    byte[] tableName = Bytes.toBytes("test-table");
    int nputs = 1000000;

    if (args.length >= 1) {
      tableName = Bytes.toBytes(args[0]);
    }

    if (args.length >= 2) {
      nputs = Long.parseLong(args[1]);
    }

    byte[] cf0 = Bytes.toBytes("cf0");
    byte[] cf1 = Bytes.toBytes("cf1");

    Configuration conf = HBaseConfiguration.create();
    HBaseAdmin admin = new HBaseAdmin(conf);
    try {
      if (!admin.tableExists(tableName)) {
        createTable(admin, tableName, cf0, cf1);
      }
      HTable table = new HTable(conf, tableName);
      loadData(table, nputs, cf0, cf1);
      admin.flush(tableName);
    } finally {
      admin.close();
    }
  }
}