Exemplo n.º 1
0
        /// <exception cref="System.Exception"/>
        protected internal virtual void _testMOWithJavaSerialization(bool withCounters)
        {
            string        input = "a\nb\nc\nd\ne\nc\nd\ne";
            Configuration conf  = CreateJobConf();

            conf.Set("io.serializations", "org.apache.hadoop.io.serializer.JavaSerialization,"
                     + "org.apache.hadoop.io.serializer.WritableSerialization");
            Job job = MapReduceTestUtil.CreateJob(conf, InDir, OutDir, 2, 1, input);

            job.SetJobName("mo");
            MultipleOutputs.AddNamedOutput(job, Text, typeof(TextOutputFormat), typeof(long),
                                           typeof(string));
            MultipleOutputs.SetCountersEnabled(job, withCounters);
            job.SetSortComparatorClass(typeof(JavaSerializationComparator));
            job.SetMapOutputKeyClass(typeof(long));
            job.SetMapOutputValueClass(typeof(string));
            job.SetOutputKeyClass(typeof(long));
            job.SetOutputValueClass(typeof(string));
            job.SetMapperClass(typeof(TestMRMultipleOutputs.MOJavaSerDeMap));
            job.SetReducerClass(typeof(TestMRMultipleOutputs.MOJavaSerDeReduce));
            job.WaitForCompletion(true);
            // assert number of named output part files
            int        namedOutputCount      = 0;
            int        valueBasedOutputCount = 0;
            FileSystem fs = OutDir.GetFileSystem(conf);

            FileStatus[] statuses = fs.ListStatus(OutDir);
            foreach (FileStatus status in statuses)
            {
                string fileName = status.GetPath().GetName();
                if (fileName.Equals("text-m-00000") || fileName.Equals("text-m-00001") || fileName
                    .Equals("text-r-00000"))
                {
                    namedOutputCount++;
                }
                else
                {
                    if (fileName.Equals("a-r-00000") || fileName.Equals("b-r-00000") || fileName.Equals
                            ("c-r-00000") || fileName.Equals("d-r-00000") || fileName.Equals("e-r-00000"))
                    {
                        valueBasedOutputCount++;
                    }
                }
            }
            NUnit.Framework.Assert.AreEqual(3, namedOutputCount);
            NUnit.Framework.Assert.AreEqual(5, valueBasedOutputCount);
            // assert TextOutputFormat files correctness
            BufferedReader reader = new BufferedReader(new InputStreamReader(fs.Open(new Path
                                                                                         (FileOutputFormat.GetOutputPath(job), "text-r-00000"))));
            int    count = 0;
            string line  = reader.ReadLine();

            while (line != null)
            {
                NUnit.Framework.Assert.IsTrue(line.EndsWith(Text));
                line = reader.ReadLine();
                count++;
            }
            reader.Close();
            NUnit.Framework.Assert.IsFalse(count == 0);
            if (withCounters)
            {
                CounterGroup counters = job.GetCounters().GetGroup(typeof(MultipleOutputs).FullName
                                                                   );
                NUnit.Framework.Assert.AreEqual(6, counters.Size());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter(Text).GetValue());
                NUnit.Framework.Assert.AreEqual(2, counters.FindCounter("a").GetValue());
                NUnit.Framework.Assert.AreEqual(2, counters.FindCounter("b").GetValue());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter("c").GetValue());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter("d").GetValue());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter("e").GetValue());
            }
        }
Exemplo n.º 2
0
 protected override void Setup(Mapper.Context context)
 {
     mos = new MultipleOutputs(context);
 }
Exemplo n.º 3
0
 protected override void Setup(Reducer.Context context)
 {
     mos = new MultipleOutputs <long, string>(context);
 }
Exemplo n.º 4
0
        /// <exception cref="System.Exception"/>
        protected internal virtual void _testMultipleOutputs(bool withCounters)
        {
            string        input = "a\nb\nc\nd\ne\nc\nd\ne";
            Configuration conf  = CreateJobConf();
            Job           job   = MapReduceTestUtil.CreateJob(conf, InDir, OutDir, 2, 1, input);

            job.SetJobName("mo");
            MultipleOutputs.AddNamedOutput(job, Text, typeof(TextOutputFormat), typeof(LongWritable
                                                                                       ), typeof(Text));
            MultipleOutputs.AddNamedOutput(job, Sequence, typeof(SequenceFileOutputFormat), typeof(
                                               IntWritable), typeof(Text));
            MultipleOutputs.SetCountersEnabled(job, withCounters);
            job.SetMapperClass(typeof(TestMRMultipleOutputs.MOMap));
            job.SetReducerClass(typeof(TestMRMultipleOutputs.MOReduce));
            job.WaitForCompletion(true);
            // assert number of named output part files
            int        namedOutputCount      = 0;
            int        valueBasedOutputCount = 0;
            FileSystem fs = OutDir.GetFileSystem(conf);

            FileStatus[] statuses = fs.ListStatus(OutDir);
            foreach (FileStatus status in statuses)
            {
                string fileName = status.GetPath().GetName();
                if (fileName.Equals("text-m-00000") || fileName.Equals("text-m-00001") || fileName
                    .Equals("text-r-00000") || fileName.Equals("sequence_A-m-00000") || fileName.Equals
                        ("sequence_A-m-00001") || fileName.Equals("sequence_B-m-00000") || fileName.Equals
                        ("sequence_B-m-00001") || fileName.Equals("sequence_B-r-00000") || fileName.Equals
                        ("sequence_C-r-00000"))
                {
                    namedOutputCount++;
                }
                else
                {
                    if (fileName.Equals("a-r-00000") || fileName.Equals("b-r-00000") || fileName.Equals
                            ("c-r-00000") || fileName.Equals("d-r-00000") || fileName.Equals("e-r-00000"))
                    {
                        valueBasedOutputCount++;
                    }
                }
            }
            NUnit.Framework.Assert.AreEqual(9, namedOutputCount);
            NUnit.Framework.Assert.AreEqual(5, valueBasedOutputCount);
            // assert TextOutputFormat files correctness
            BufferedReader reader = new BufferedReader(new InputStreamReader(fs.Open(new Path
                                                                                         (FileOutputFormat.GetOutputPath(job), "text-r-00000"))));
            int    count = 0;
            string line  = reader.ReadLine();

            while (line != null)
            {
                NUnit.Framework.Assert.IsTrue(line.EndsWith(Text));
                line = reader.ReadLine();
                count++;
            }
            reader.Close();
            NUnit.Framework.Assert.IsFalse(count == 0);
            // assert SequenceOutputFormat files correctness
            SequenceFile.Reader seqReader = new SequenceFile.Reader(fs, new Path(FileOutputFormat
                                                                                 .GetOutputPath(job), "sequence_B-r-00000"), conf);
            NUnit.Framework.Assert.AreEqual(typeof(IntWritable), seqReader.GetKeyClass());
            NUnit.Framework.Assert.AreEqual(typeof(Text), seqReader.GetValueClass());
            count = 0;
            IntWritable key   = new IntWritable();
            Text        value = new Text();

            while (seqReader.Next(key, value))
            {
                NUnit.Framework.Assert.AreEqual(Sequence, value.ToString());
                count++;
            }
            seqReader.Close();
            NUnit.Framework.Assert.IsFalse(count == 0);
            if (withCounters)
            {
                CounterGroup counters = job.GetCounters().GetGroup(typeof(MultipleOutputs).FullName
                                                                   );
                NUnit.Framework.Assert.AreEqual(9, counters.Size());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter(Text).GetValue());
                NUnit.Framework.Assert.AreEqual(2, counters.FindCounter(Sequence + "_A").GetValue
                                                    ());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter(Sequence + "_B").GetValue
                                                    ());
                NUnit.Framework.Assert.AreEqual(2, counters.FindCounter(Sequence + "_C").GetValue
                                                    ());
                NUnit.Framework.Assert.AreEqual(2, counters.FindCounter("a").GetValue());
                NUnit.Framework.Assert.AreEqual(2, counters.FindCounter("b").GetValue());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter("c").GetValue());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter("d").GetValue());
                NUnit.Framework.Assert.AreEqual(4, counters.FindCounter("e").GetValue());
            }
        }