HBase 与 MapReduce 集成

6. HBase 与 MapReduce 集成

6.1 官方 HBase 与 MapReduce 集成

查看 HBase 的 MapReduce 任务的执行:bin/hbase mapredcp;
环境变量的导入
1. 临时生效,在命令行执行操作:
  - export HBASE_HOME=/opt/module/hbase-1.3.4;
  - export HADOOP_HOME=/opt/module/hadoop-2.8.5;
  - export HADOOP_CLASSPATH=${HBASE_HOME}/bin/hbase mapredcp;
2. 永久生效,在/etc/profile配置
  - export HBASE_HOME=/opt/module/hbase-1.3.4;
  - export HADOOP_HOME=/opt/module/hadoop-2.8.5;
  - 并在hadoop-env.sh配置:export HADOOP_CLASSPATH=$HADOOP_CLASSPATH:/opt/module/hbase/lib/*
运行官方的 MapReduce 任务

// ===== 案例一:统计Student表中有多少行数据 (`opt/module/hbase-1.3.4/` 目录下)
/opt/module/hadoop-2.8.5/bin/yarn jar ./lib/hbase-server-1.3.4.jar rowcounter student


// ===== 案例二:使用 MapReduce 将本地数据导入到 HBASE
// 1. 本地创建一个fruit.tsv文件
1001    Apple   Red
1002    Pear    Yellow
1003    Pineapple   Yellow

// 2. 创建 HBase 表
create 'fruit','info'

// 3. 在 HDFS 中创建 input_fruit 文件夹并上传 fruit.tsv 文件
/opt/module/hadoop-2.8.5/bin/hdfs dfs -mkdir /input_fruit
/opt/module/hadoop-2.8.5/bin/hdfs dfs -put fruit.tsv /input_fruit/

// 4. 执行 MapReduce, 将 fruit.tsv 导入到 HBase 的 fruit 表中
/opt/module/hadoop-2.8.5/bin/yarn jar ./lib/hbase-server-1.3.4.jar importtsv -Dimporttsv.columns=HBASE_ROW_KEY,info:name,info:color fruit hdfs://IP地址/input_fruit

6.2 自定义HBase-MapReduce

需求:将 fruit 表中的部分数据,通过MR迁入到 fruit_mr 表中

// 1. 创建 FruitMapper 类,用于读取 fruit 表中的数据
public class FruitMapper extends TableMapper<ImmutableBytesWritable, Put>{

	@Override
	protected void map(ImmutableBytesWritable key, Result value, Context context) throws IOException, InterruptedException {
		// 创建put对象
		Put put = new Put(key.get());
		
		Cell[] cells = value.rawCells();
		
		for(Cell cell : cells) {
			if("name".equals(Bytes.toString(CellUtil.cloneQualifier(cell)))) {
				put.add(cell);
			}
		}
		
		context.write(key, put);
	}
}

// 2. 创建 FruitReducer 类,用于写入 
public class FruitReducer extends TableReducer<ImmutableBytesWritable, Put, NullWritable>{

	@Override
	protected void reduce(ImmutableBytesWritable key, Iterable<Put> values, Context context) throws IOException, InterruptedException {
		for (Put value : values) {
			context.write(NullWritable.get(), value);
		}
	}
}

// 3. 创建 FruitDriver 类,用于执行 mapper 和 reducer
public class FruitDriver extends Configuration implements Tool{

	private Configuration configuration = null;
	
	@Override
	public void setConf(Configuration conf) {
		this.configuration = conf;
	}
	
	@Override
	public Configuration getConf() {
		return configuration;
	}

	@Override
	public int run(String[] args) throws Exception {
		// 获取任务对象
		Job job = Job.getInstance(configuration);
		
		// 指定 Driver类
		job.setJarByClass(FruitDriver.class);
		
		// 指定 Mapper
		TableMapReduceUtil.initTableMapperJob("fruit", new Scan(), FruitMapper.class, ImmutableBytesWritable.class, Put.class, job);
		
		// 指定 Reducer
		TableMapReduceUtil.initTableReducerJob("fruit_mr", FruitReducer.class, job);
		
		// 提交
		boolean result = job.waitForCompletion(true);
		
		return result ? 0 : 1;
	}

	public static void main(String[] args) throws Exception {
		
		Configuration configuration = HBaseConfiguration.create();
		ToolRunner.run(configuration, new FruitDriver(), args);
	}
}

// 4. 打成 fruit.jar包
// 5. HBase 中创建 fruit_mr 表
create 'fruit_mr','info'

// 6. 在 /opt/module/hbase 中执行:
/opt/module/hadoop-2.8.5/bin/yarn jar ./fruit.jar com.noodles.mr1.FruitDriver(Driver的类名)

6.3 自定义 HBase-MapReduce2

需求:实现将 HDFS 中的数据写入到 HBase 表中

// 1. 创建 Mapper, 用于读取 HDFS 上的文件
public class HDFSMapper extends Mapper<LongWritable, Text, NullWritable, Put>{

	@Override
	protected void map(LongWritable key, Text value, Mapper<LongWritable, Text, NullWritable, Put>.Context context)
			throws IOException, InterruptedException {
		// 获取一行数据
		String line = value.toString();
		
		// 切割
		String[] split = line.split("	");
		
		// 封装 Put 对象
		Put put = new Put(Bytes.toBytes(split[0]));
		put.addColumn(Bytes.toBytes("info"), Bytes.toBytes("name"), Bytes.toBytes(split[1]));
		put.addColumn(Bytes.toBytes("info"), Bytes.toBytes("color"), Bytes.toBytes(split[2]));
		
		// 写出去
		context.write(NullWritable.get(), put);
	}
}

// 2. 创建 Reducer, 用于写入
public class HDFSReducer extends TableReducer<NullWritable, Put, NullWritable>{

	@Override
	protected void reduce(NullWritable key, Iterable<Put> values,
			Reducer<NullWritable, Put, NullWritable, Mutation>.Context context) throws IOException, InterruptedException {
		
		// 写出数据
		for(Put value : values) {
			context.write(NullWritable.get(), value);
		}
	}
}

// 3. 创建Driver
public class HDFSDriver extends Configuration implements Tool{
	
	private Configuration configuration = null;

	@Override
	public void setConf(Configuration conf) {
		this.configuration = conf;
	}

	@Override
	public Configuration getConf() {
		return configuration;
	}

	@Override
	public int run(String[] args) throws Exception {

		// 获取 Job 对象
		Job job = Job.getInstance(configuration);
		
		// 设置主类
		job.setJarByClass(HDFSDriver.class);
		
		// 设置 Mapper
		job.setMapperClass(HDFSMapper.class);
		job.setMapOutputKeyClass(NullWritable.class);
		job.setMapOutputValueClass(Put.class);
		
		// 设置 Reducer
		TableMapReduceUtil.initTableReducerJob("fruit2", HDFSReducer.class, job);

        // 设置输入路径
		// import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
		FileInputFormat.setInputPaths(job, args[0]);
		
		// 提交
		boolean result = job.waitForCompletion(true);
		
		return result ? 0 : 1;
	}
	
	public static void main(String[] args) throws Exception {
		
		Configuration configuration = HBaseConfiguration.create();
		ToolRunner.run(configuration, new HDFSDriver(), args);

	}
}

// 4. 打成 fruit.jar包
// 5. HBase 中创建 fruit2 表
create 'fruit2','info'

// 6. 在 /opt/module/hbase 中执行:
/opt/module/hadoop-2.8.5/bin/yarn jar ./fruit.jar com.noodles.mr2.HDFSDriver(Driver的类名) /input_fruit/fruit.tsv(文件路径)