Since you know the format of your output, write something to read from it should be very easy. What's your outputformat? If you didn't set it, please refer to the class "org.apache.hadoop.mapreduce.lib.output.TextOutputFormat". you can use the class "org.apache.hadoop.mapreduce.lib.input.TextInputFormat" to read each record from you output file.
Regards, *Stanley Shi,* On Tue, Mar 25, 2014 at 9:07 PM, Li Li <fancye...@gmail.com> wrote: > I have map-reduce job to output my custom writable objects, how can I > read it using pure java api? > I don't want to serialize it to string(Text) and deserialize using java. > > map reduce codes.. > .. > job.setOutputKeyClass(IntWritable.class); > job.setOutputValueClass(MyWritable.class); >