hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Becker <_martinbec...@web.de>
Subject Re: 0.21.0 API
Date Wed, 22 Sep 2010 15:45:31 GMT
Tom, thanks for your answer.

OK, so the problem is, that when I add both hadoop-common-0.21.0.jar AND 
hadoop-mapred-0.21.0.jar to the class path, the Tool interface is marked 
deprecated by Eclipse. This seems odd. Only having 
hadoop-common-0.21.0.jar in the class path works fine. Both jars define 
classes in the package org.apache.hadoop.util where the Tool interface 
is located. Could this be the problem? I am a little clueless here and 
not sure whether this is a problem that should be further addressed in 
this mailing list.

Thanks in advance,

On 22.09.2010 16:08, Tom White wrote:
> Hi Martin,
> Neither Tool nor ToolRunner is deprecated in 0.21.0. I don't think
> they have ever been deprecated. You should be able to use them without
> problems.
> Tom
> On Wed, Sep 22, 2010 at 6:52 AM, Martin Becker<_martinbecker@web.de>  wrote:
>> Hello,
>> I am trying to move to Hadoop MapReduce 0.21.0.
>> The corresponding tutorial still uses Tool and ToolRunner.
>> Yet both are deprecated. What would be the correct way to implement,
>> configure and submit a Job now? I was thinking in terms of:
>>      Configuration configuration = new Configuration();         Cluster
>> cluster = new Cluster(configuration);     Job job =
>> Job.getInstance(cluster);
>>      job.setJarByClass(WordCount.class);     job.setMapperClass(Map.class);
>>      job.setCombinerClass(Reduce.class);
>> job.setReducerClass(Reduce.class);     job.setOutputKeyClass(Text.class);
>>      job.setOutputValueClass(IntWritable.class);
>>      FileInputFormat.addInputPath(job, new Path(INPUT));
>> FileOutputFormat.setOutputPath(job, new Path(OUTPUT));
>>      System.exit(job.waitForCompletion(true) ? 0 : 1); Thanks in advance,
>> Martin

View raw message