Hbase 与mapreduce结合

Hbase和mapreduce结合

为什么需要用mapreduce去访问hbase的数据?

——加快分析速度和扩展分析能力

Mapreduce访问hbase数据作分析一定是在离线分析的场景下应用

 

案例1、HBase表数据的转移

在Hadoop阶段,我们编写的MR任务分别进程了Mapper和Reducer两个类,而在HBase中我们需要继承的是TableMapper和TableReducer两个类。

目标:将fruit表中的一部分数据,通过MR迁入到fruit_mr表中

Step1、构建ReadFruitMapper类,用于读取fruit表中的数据


import java.io.IOException;

import org.apache.hadoop.hbase.Cell;

import org.apache.hadoop.hbase.CellUtil;

import org.apache.hadoop.hbase.client.Put;

import org.apache.hadoop.hbase.client.Result;

import org.apache.hadoop.hbase.io.ImmutableBytesWritable;

import org.apache.hadoop.hbase.mapreduce.TableMapper;

import org.apache.hadoop.hbase.util.Bytes;

public class ReadFruitMapper extends TableMapper<ImmutableBytesWritable, Put> {

       @Override

       protected void map(ImmutableBytesWritable key, Result value, Context context)

       throws IOException, InterruptedException {

              //将fruit的name和color提取出来,相当于将每一行数据读取出来放入到Put对象中。

              Put put = new Put(key.get());

              //遍历添加column行

              for(Cell cell: value.rawCells()){

                     //添加/克隆列族:info

                     if("info".equals(Bytes.toString(CellUtil.cloneFamily(cell)))){

                            //添加/克隆列:name

                            if("name".equals(Bytes.toString(CellUtil.cloneQualifier(cell)))){

                                   //将该列cell加入到put对象中

                                   put.add(cell);

                                   //添加/克隆列:color

                            }else if("color".equals(Bytes.toString(CellUtil.cloneQualifier(cell)))){

                                   //向该列cell加入到put对象中

                                   put.add(cell);

                            }

                     }

              }

              //将从fruit读取到的每行数据写入到context中作为map的输出

              context.write(key, put);

       }

}


Step2、构建WriteFruitMRReducer类,用于将读取到的fruit表中的数据写入到fruit_mr表中


import java.io.IOException;

import org.apache.hadoop.hbase.client.Put;

import org.apache.hadoop.hbase.io.ImmutableBytesWritable;

import org.apache.hadoop.hbase.mapreduce.TableReducer;

import org.apache.hadoop.io.NullWritable;

public class WriteFruitMRReducer extends TableReducer<ImmutableBytesWritable, Put, NullWritable> {

       @Override

       protected void reduce(ImmutableBytesWritable key, Iterable<Put> values, Context context)

       throws IOException, InterruptedException {

              //读出来的每一行数据写入到fruit_mr表中

              for(Put put: values){

                     context.write(NullWritable.get(), put);

              }

       }

      

}


Step3、构建Fruit2FruitMRJob extends Configured implements Tool,用于组装运行Job任务

   


//组装Job

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

              //得到Configuration

              Configuration conf = this.getConf();

              //创建Job任务

              Job job = Job.getInstance(conf, this.getClass().getSimpleName());

              job.setJarByClass(Fruit2FruitMRJob.class);

              //配置Job

              Scan scan = new Scan();

              scan.setCacheBlocks(false);

              scan.setCaching(500);

              //设置Mapper,注意导入的是mapreduce包下的,不是mapred包下的,后者是老版本

              TableMapReduceUtil.initTableMapperJob(

              "fruit", //数据源的表名

              scan, //scan扫描控制器

              ReadFruitMapper.class,//设置Mapper类

              ImmutableBytesWritable.class,//设置Mapper输出key类型

              Put.class,//设置Mapper输出value值类型

              job//设置给哪个JOB

              );

              //设置Reducer

              TableMapReduceUtil.initTableReducerJob("fruit_mr", WriteFruitMRReducer.class, job);

              //设置Reduce数量,最少1个

              job.setNumReduceTasks(1);

              boolean isSuccess = job.waitForCompletion(true);

              if(!isSuccess){

                     throw new IOException("Job running with error");

              }

             

              return isSuccess ? 0 : 1;

       }


Step4、主函数中调用运行该Job任务


public static void main( String[] args ) throws Exception{

Configuration conf = HBaseConfiguration.create();

int status = ToolRunner.run(conf, new Fruit2FruitMRJob(), args);

System.exit(status);

}


 

案例2:从Hbase中读取数据、分析,写入hdfs

/**

public abstract class TableMapper<KEYOUT, VALUEOUT>

extends Mapper<ImmutableBytesWritable, Result, KEYOUT, VALUEOUT> {

}

 * @author duanhaitao@gec.cn

 *

 */

public class HbaseReader {

 

       public static String flow_fields_import = "flow_fields_import";

       static class HdfsSinkMapper extends TableMapper<Text, NullWritable>{

 

              @Override

              protected void map(ImmutableBytesWritable key, Result value, Context context) throws IOException, InterruptedException {

 

                     byte[] bytes = key.copyBytes();

                     String phone = new String(bytes);

                     byte[] urlbytes = value.getValue("f1".getBytes(), "url".getBytes());

                     String url = new String(urlbytes);

                     context.write(new Text(phone + " " + url), NullWritable.get());

                    

              }

             

       }

      

       static class HdfsSinkReducer extends Reducer<Text, NullWritable, Text, NullWritable>{

             

              @Override

              protected void reduce(Text key, Iterable<NullWritable> values, Context context) throws IOException, InterruptedException {

                    

                     context.write(key, NullWritable.get());

              }

       }

      

       public static void main(String[] args) throws Exception {

              Configuration conf = HBaseConfiguration.create();

              conf.set("hbase.zookeeper.quorum", "spark01");

             

              Job job = Job.getInstance(conf);

             

              job.setJarByClass(HbaseReader.class);

             

//            job.setMapperClass(HdfsSinkMapper.class);

              Scan scan = new Scan();

              TableMapReduceUtil.initTableMapperJob(flow_fields_import, scan, HdfsSinkMapper.class, Text.class, NullWritable.class, job);

              job.setReducerClass(HdfsSinkReducer.class);

             

              FileOutputFormat.setOutputPath(job, new Path("c:/hbasetest/output"));

             

              job.setOutputKeyClass(Text.class);

              job.setOutputValueClass(NullWritable.class);

             

              job.waitForCompletion(true);

       }

      

}

2.3.2 从hdfs中读取数据写入Hbase

q

/**

public abstract class TableReducer<KEYIN, VALUEIN, KEYOUT>

extends Reducer<KEYIN, VALUEIN, KEYOUT, Writable> {

}

 * @author duanhaitao@gec.cn

 *

 */

public class HbaseSinker {

 

       public static String flow_fields_import = "flow_fields_import";

       static class HbaseSinkMrMapper extends Mapper<LongWritable, Text, FlowBean, NullWritable>{

              @Override

              protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

 

                     String line = value.toString();

                     String[] fields = line.split(" ");

                     String phone = fields[0];

                     String url = fields[1];

                    

                     FlowBean bean = new FlowBean(phone,url);

                    

                     context.write(bean, NullWritable.get());

              }

       }

      

       static class HbaseSinkMrReducer extends TableReducer<FlowBean, NullWritable, ImmutableBytesWritable>{

             

              @Override

              protected void reduce(FlowBean key, Iterable<NullWritable> values, Context context) throws IOException, InterruptedException {

                    

                     Put put = new Put(key.getPhone().getBytes());

                     put.add("f1".getBytes(), "url".getBytes(), key.getUrl().getBytes());

                    

                     context.write(new ImmutableBytesWritable(key.getPhone().getBytes()), put);

                    

              }

             

       }

      

       public static void main(String[] args) throws Exception {

              Configuration conf = HBaseConfiguration.create();

              conf.set("hbase.zookeeper.quorum", "spark01");

             

              HBaseAdmin hBaseAdmin = new HBaseAdmin(conf);

             

              boolean tableExists = hBaseAdmin.tableExists(flow_fields_import);

              if(tableExists){

                     hBaseAdmin.disableTable(flow_fields_import);

                     hBaseAdmin.deleteTable(flow_fields_import);

              }

              HTableDescriptor desc = new HTableDescriptor(TableName.valueOf(flow_fields_import));

              HColumnDescriptor hColumnDescriptor = new HColumnDescriptor ("f1".getBytes());

              desc.addFamily(hColumnDescriptor);

             

              hBaseAdmin.createTable(desc);

             

             

              Job job = Job.getInstance(conf);

             

              job.setJarByClass(HbaseSinker.class);

             

              job.setMapperClass(HbaseSinkMrMapper.class);

              TableMapReduceUtil.initTableReducerJob(flow_fields_import, HbaseSinkMrReducer.class, job);

             

              FileInputFormat.setInputPaths(job, new Path("c:/hbasetest/data"));

             

              job.setMapOutputKeyClass(FlowBean.class);

              job.setMapOutputValueClass(NullWritable.class);

             

              job.setOutputKeyClass(ImmutableBytesWritable.class);

              job.setOutputValueClass(Mutation.class);

             

              job.waitForCompletion(true);

             

             

       }

      

}

原文地址:https://www.cnblogs.com/Transkai/p/10748536.html