avro-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Something Something <mailinglist...@gmail.com>
Subject Re: Reading AVRO files in Java MapReduce
Date Thu, 17 Nov 2011 00:00:12 GMT
Why could this not be working?  Any ideas?  I even put a 'throw new
RuntimeException' to see if it's coming to the Mapper, but it isn't.
 Thanks for the help.

Code snippet:



    public static class MapImpl extends AvroMapper<Utf8, Pair<Utf8, Long>> {

        @Override

        public void map(Utf8 text, AvroCollector<Pair<Utf8, Long>>
collector,

                        Reporter reporter) throws IOException {

        throw new RuntimeException("my text: " + text.toString());

//        System.out.println("my text" + text);

//

//            collector.collect(new Pair<Utf8, Long>(text, 1L));

        }

    }



    private static class NonAvroReducer

            extends MapReduceBase

            implements Reducer<AvroKey<Utf8>, AvroValue<Long>, Text, Text>
{


        public void reduce(AvroKey<Utf8> key, Iterator<AvroValue<Long>>
values,

                           OutputCollector<Text, Text> out,

                           Reporter reporter) throws IOException {

        out.collect(new Text(key.toString()),

                    new Text("Testing"));

            while (values.hasNext()) {

                AvroValue<Long> value = values.next();

                out.collect(new Text(key.toString()),

                        new Text(value.datum().toString()));

            }

        }

    }


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



        String dir = "/user/mydir";



        JobConf job = new JobConf(new Configuration(), TestAvroProcessor.
class);

        job.setJobName(TestAvroProcessor.class.getName());



        Path outputPath = new Path(dir + "/out");


        outputPath.getFileSystem(job).delete(outputPath);



        AvroJob.setInputSchema(job, Schema.parse(new File(
"/Users/mydir/profiles.json")));

        AvroJob.setOutputSchema(job, SCHEMA);


        AvroJob.setMapperClass(job, MapImpl.class);


        FileInputFormat.setInputPaths(job, new Path(dir + "/data"));

        FileOutputFormat.setOutputPath(job, outputPath);

        FileOutputFormat.setCompressOutput(job, false);



        job.setReducerClass(NonAvroReducer.class);

        job.setOutputFormat(TextOutputFormat.class);

        job.setOutputKeyClass(Text.class);

        job.setOutputValueClass(Text.class);



        JobClient.runJob(job);





    }




On Tue, Nov 15, 2011 at 5:12 PM, Doug Cutting <cutting@apache.org> wrote:

> On 11/15/2011 03:16 PM, Something Something wrote:
> > Quick question.  I want the output from AvroJob.setReducerClass to be in
> > regular Text files - not in AVRO format.  Can I do that?  Any examples?
> >  Sorry, kinda short on time to do research.  Thanks.
>
> On the previously cited documentation page:
>
>
> http://avro.apache.org/docs/current/api/java/org/apache/avro/mapred/package-summary.html
>
> Look for the text, "For jobs whose input is an Avro data file and which
> use an AvroMapper, but whose reducer is a non-Avro Reducer and whose
> output is a non-Avro format".
>
> A sample of a job that does this is at:
>
>  http://s.apache.org/MsG
>
> Just use TextOutputFormat instead of SequenceFileOutputFormat.
>
> Doug
>

Mime
View raw message