Mongo-Hadoop

下载

https://github.com/mongodb/mongo-hadoop/releases

 

解压到/home/kevin/hadoop/hadoop/share/mongo-hadoop-2.2-1.2.0

（其实是任意，真正分布式的时候要注意这个jar包的问题）

 

Mapreduce结果输出到MongDB

要修改的代码（input类似  ps:extends Mapper<Object,BSONObject, IntWritable,DoubleWritable>）

//FileOutputFormat.setOutputPath(job, new Path(otherArgs[1]));
MongoConfigUtil.setOutputURI(conf, "mongodb://localhost:27017/test.out");
 job.setOutputFormatClass(MongoOutputFormat.class);

 

注意：

import的不要是 com.mongodb.hadoop.mapred.MongoInputFormat;，而是

import com.mongodb.hadoop.MongoOutputFormat;
import com.mongodb.hadoop.MongoInputFormat;
import com.mongodb.hadoop.util.MongoConfigUtil;

 

　　　BasicBSONObject output = new BasicBSONObject ();

         output . put ( "count" , count );

         output . put ( "avg" , avg );

         output . put ( "sum" , sum );

         pContext . write ( pKey , new BSONWritable ( output ));