您好,登錄后才能下訂單哦!
這篇文章主要為大家展示了“hbase如何編寫mapreduce”,內容簡而易懂,條理清晰,希望能夠幫助大家解決疑惑,下面讓小編帶領大家一起研究并學習一下“hbase如何編寫mapreduce”這篇文章吧。
package com.hbase.test; import java.io.IOException; import org.apache.hadoop.conf.Configuration; import org.apache.hadoop.hbase.HBaseConfiguration; import org.apache.hadoop.hbase.client.Mutation; import org.apache.hadoop.hbase.client.Put; import org.apache.hadoop.hbase.client.Result; import org.apache.hadoop.hbase.client.Scan; import org.apache.hadoop.hbase.io.ImmutableBytesWritable; import org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil; import org.apache.hadoop.hbase.mapreduce.TableMapper; import org.apache.hadoop.hbase.mapreduce.TableReducer; import org.apache.hadoop.hbase.util.Bytes; import org.apache.hadoop.io.IntWritable; import org.apache.hadoop.io.Text; import org.apache.hadoop.mapreduce.Job; import org.apache.hadoop.mapreduce.Mapper; import org.apache.hadoop.mapreduce.Reducer; public class HbaseMrTest { public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException { Configuration conf = HBaseConfiguration.create(); //配置conf conf.set("hbase.zookeeper.quorum", "bigdata01,bigdata02,bigdata03"); conf.set("hbase.zookeeper.property.clientPort", "2181"); Job job = Job.getInstance(conf, "word-count"); //指定執行job的主類 job.setJarByClass(HbaseMrTest.class); Scan scan = new Scan(); //定義mapper需要掃描的列 scan.addColumn(Bytes.toBytes("content"), Bytes.toBytes("words")); //配置mapper TableMapReduceUtil.initTableMapperJob("wordcount", scan,HMapper.class , Text.class, IntWritable.class, job); //配置recuder TableMapReduceUtil.initTableReducerJob("result", HReducer.class, job); //提交job System.exit(job.waitForCompletion(true)?0:1); } } // Text, IntWritable 為輸出類型 class HMapper extends TableMapper<Text, IntWritable>{ Text out = new Text(); IntWritable iw = new IntWritable(1); @Override protected void map(ImmutableBytesWritable key, Result value, Mapper<ImmutableBytesWritable, Result, Text, IntWritable>.Context context) throws IOException, InterruptedException { //通過result 直接過得content:words 的值 byte[] bytes = value.getValue(Bytes.toBytes("content"), Bytes.toBytes("words")); if(bytes!=null) { String words = Bytes.toString(bytes); //對獲得的一行單詞進行分割 String[] ws = words.split(" "); for(String wd : ws) { out.set(wd); //寫出值,如: you 1 context.write(out, iw); } } } } // Text, IntWritable 為mapper的輸出類型 class HReducer extends TableReducer<Text, IntWritable, ImmutableBytesWritable>{ @Override protected void reduce(Text text, Iterable<IntWritable> iter, Reducer<Text, IntWritable, ImmutableBytesWritable, Mutation>.Context context) throws IOException, InterruptedException { int sum = 0 ; //對iter遍歷 for(IntWritable intw : iter) { sum+= intw.get(); } //new 一個put 構造函數內的值為row key Put put = new Put(Bytes.toBytes(text.toString())); //put添加columnfamily 和column put.addColumn(Bytes.toBytes("info"), Bytes.toBytes("wordcnt"), Bytes.toBytes(String.valueOf(sum))); //將每個單詞當做row key 寫出,put是相加的總和 context.write(new ImmutableBytesWritable(Bytes.toBytes(text.toString())), put); } } 最后將java文件export為RaunableJar放到linux java -jar hbase.jar com.hbase.test.HbaseMrTest 運行
原始數據:
運行結果:
以上是“hbase如何編寫mapreduce”這篇文章的所有內容,感謝各位的閱讀!相信大家都有了一定的了解,希望分享的內容對大家有所幫助,如果還想學習更多知識,歡迎關注億速云行業資訊頻道!
免責聲明:本站發布的內容(圖片、視頻和文字)以原創、轉載和分享為主,文章觀點不代表本網站立場,如果涉及侵權請聯系站長郵箱:is@yisu.com進行舉報,并提供相關證據,一經查實,將立刻刪除涉嫌侵權內容。