internal BinarySearchNode(TotalOrderPartitioner <K, V> _enclosing, K[] splitPoints
, RawComparator <K> comparator)
{
this._enclosing = _enclosing;
this.splitPoints = splitPoints;
this.comparator = comparator;
}
internal UnsplitTrieNode(TotalOrderPartitioner <K, V> _enclosing, int level, int value
)
: base(level)
{
this._enclosing = _enclosing;
this.result = value;
}
/// <exception cref="System.IO.IOException"/>
private static Path WritePartitionFile <T>(string testname, Configuration conf, T[]
splits)
where T : WritableComparable <object>
{
FileSystem fs = FileSystem.GetLocal(conf);
Path testdir = new Path(Runtime.GetProperty("test.build.data", "/tmp")).MakeQualified
(fs);
Path p = new Path(testdir, testname + "/_partition.lst");
TotalOrderPartitioner.SetPartitionFile(conf, p);
conf.SetInt(MRJobConfig.NumReduces, splits.Length + 1);
SequenceFile.Writer w = null;
try
{
w = SequenceFile.CreateWriter(fs, conf, p, splits[0].GetType(), typeof(NullWritable
), SequenceFile.CompressionType.None);
for (int i = 0; i < splits.Length; ++i)
{
w.Append(splits[i], NullWritable.Get());
}
}
finally
{
if (null != w)
{
w.Close();
}
}
return(p);
}
/// <exception cref="System.Exception"/>
public virtual void TestTotalOrderBinarySearch()
{
TotalOrderPartitioner <Text, NullWritable> partitioner = new TotalOrderPartitioner
<Text, NullWritable>();
Configuration conf = new Configuration();
Path p = TestTotalOrderPartitioner.WritePartitionFile <Text>("totalorderbinarysearch"
, conf, splitStrings);
conf.SetBoolean(TotalOrderPartitioner.NaturalOrder, false);
conf.SetClass(MRJobConfig.MapOutputKeyClass, typeof(Text), typeof(object));
try
{
partitioner.SetConf(conf);
NullWritable nw = NullWritable.Get();
foreach (TestTotalOrderPartitioner.Check <Text> chk in testStrings)
{
NUnit.Framework.Assert.AreEqual(chk.data.ToString(), chk.part, partitioner.GetPartition
(chk.data, nw, splitStrings.Length + 1));
}
}
finally
{
p.GetFileSystem(conf).Delete(p, true);
}
}
internal SinglySplitTrieNode(TotalOrderPartitioner <K, V> _enclosing, int level, BinaryComparable
[] splitPoints, int lower)
: base(level)
{
this._enclosing = _enclosing;
this.lower = lower;
this.mySplitPoint = splitPoints[lower];
}
internal LeafTrieNode(TotalOrderPartitioner <K, V> _enclosing, int level, BinaryComparable
[] splitPoints, int lower, int upper)
: base(level)
{
this._enclosing = _enclosing;
this.lower = lower;
this.upper = upper;
this.splitPoints = splitPoints;
}
/// <exception cref="System.Exception"/>
public virtual void TestTotalOrderCustomComparator()
{
TotalOrderPartitioner <Text, NullWritable> partitioner = new TotalOrderPartitioner
<Text, NullWritable>();
Configuration conf = new Configuration();
Text[] revSplitStrings = Arrays.CopyOf(splitStrings, splitStrings.Length);
Arrays.Sort(revSplitStrings, new TestTotalOrderPartitioner.ReverseStringComparator
());
Path p = TestTotalOrderPartitioner.WritePartitionFile <Text>("totalordercustomcomparator"
, conf, revSplitStrings);
conf.SetBoolean(TotalOrderPartitioner.NaturalOrder, false);
conf.SetClass(MRJobConfig.MapOutputKeyClass, typeof(Text), typeof(object));
conf.SetClass(MRJobConfig.KeyComparator, typeof(TestTotalOrderPartitioner.ReverseStringComparator
), typeof(RawComparator));
AList <TestTotalOrderPartitioner.Check <Text> > revCheck = new AList <TestTotalOrderPartitioner.Check
<Text> >();
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("aaaaa"), 9));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("aaabb"), 9));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("aabbb"), 9));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("aaaaa"), 9));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("babbb"), 8));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("baabb"), 8));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("yai"), 1));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("yak"), 1));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("z"), 0));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("ddngo"), 4));
revCheck.AddItem(new TestTotalOrderPartitioner.Check <Text>(new Text("hi"), 3));
try
{
partitioner.SetConf(conf);
NullWritable nw = NullWritable.Get();
foreach (TestTotalOrderPartitioner.Check <Text> chk in revCheck)
{
NUnit.Framework.Assert.AreEqual(chk.data.ToString(), chk.part, partitioner.GetPartition
(chk.data, nw, splitStrings.Length + 1));
}
}
finally
{
p.GetFileSystem(conf).Delete(p, true);
}
}
/// <summary>Write a partition file for the given job, using the Sampler provided.</summary>
/// <remarks>
/// Write a partition file for the given job, using the Sampler provided.
/// Queries the sampler for a sample keyset, sorts by the output key
/// comparator, selects the keys for each rank, and writes to the destination
/// returned from
/// <see cref="TotalOrderPartitioner{K, V}.GetPartitionFile(Org.Apache.Hadoop.Conf.Configuration)
/// "/>
/// .
/// </remarks>
/// <exception cref="System.IO.IOException"/>
/// <exception cref="System.TypeLoadException"/>
/// <exception cref="System.Exception"/>
public static void WritePartitionFile <K, V>(Job job, InputSampler.Sampler <K, V> sampler
)
{
// getInputFormat, getOutputKeyComparator
Configuration conf = job.GetConfiguration();
InputFormat inf = ReflectionUtils.NewInstance(job.GetInputFormatClass(), conf);
int numPartitions = job.GetNumReduceTasks();
K[] samples = (K[])sampler.GetSample(inf, job);
Log.Info("Using " + samples.Length + " samples");
RawComparator <K> comparator = (RawComparator <K>)job.GetSortComparator();
Arrays.Sort(samples, comparator);
Path dst = new Path(TotalOrderPartitioner.GetPartitionFile(conf));
FileSystem fs = dst.GetFileSystem(conf);
if (fs.Exists(dst))
{
fs.Delete(dst, false);
}
SequenceFile.Writer writer = SequenceFile.CreateWriter(fs, conf, dst, job.GetMapOutputKeyClass
(), typeof(NullWritable));
NullWritable nullValue = NullWritable.Get();
float stepSize = samples.Length / (float)numPartitions;
int last = -1;
for (int i = 1; i < numPartitions; ++i)
{
int k = Math.Round(stepSize * i);
while (last >= k && comparator.Compare(samples[last], samples[k]) == 0)
{
++k;
}
writer.Append(samples[k], nullValue);
last = k;
}
writer.Close();
}
internal CarriedTrieNodeRef(TotalOrderPartitioner <K, V> _enclosing)
{
this._enclosing = _enclosing;
this.content = null;
}
internal InnerTrieNode(TotalOrderPartitioner <K, V> _enclosing, int level)
: base(level)
{
this._enclosing = _enclosing;
}
/// <summary>Driver for InputSampler from the command line.</summary>
/// <remarks>
/// Driver for InputSampler from the command line.
/// Configures a JobConf instance and calls
/// <see cref="InputSampler{K, V}.WritePartitionFile{K, V}(Org.Apache.Hadoop.Mapreduce.Job, Sampler{K, V})
/// "/>
/// .
/// </remarks>
/// <exception cref="System.Exception"/>
public virtual int Run(string[] args)
{
Job job = Job.GetInstance(GetConf());
AList <string> otherArgs = new AList <string>();
InputSampler.Sampler <K, V> sampler = null;
for (int i = 0; i < args.Length; ++i)
{
try
{
if ("-r".Equals(args[i]))
{
job.SetNumReduceTasks(System.Convert.ToInt32(args[++i]));
}
else
{
if ("-inFormat".Equals(args[i]))
{
job.SetInputFormatClass(Sharpen.Runtime.GetType(args[++i]).AsSubclass <InputFormat
>());
}
else
{
if ("-keyClass".Equals(args[i]))
{
job.SetMapOutputKeyClass(Sharpen.Runtime.GetType(args[++i]).AsSubclass <WritableComparable
>());
}
else
{
if ("-splitSample".Equals(args[i]))
{
int numSamples = System.Convert.ToInt32(args[++i]);
int maxSplits = System.Convert.ToInt32(args[++i]);
if (0 >= maxSplits)
{
maxSplits = int.MaxValue;
}
sampler = new InputSampler.SplitSampler <K, V>(numSamples, maxSplits);
}
else
{
if ("-splitRandom".Equals(args[i]))
{
double pcnt = double.ParseDouble(args[++i]);
int numSamples = System.Convert.ToInt32(args[++i]);
int maxSplits = System.Convert.ToInt32(args[++i]);
if (0 >= maxSplits)
{
maxSplits = int.MaxValue;
}
sampler = new InputSampler.RandomSampler <K, V>(pcnt, numSamples, maxSplits);
}
else
{
if ("-splitInterval".Equals(args[i]))
{
double pcnt = double.ParseDouble(args[++i]);
int maxSplits = System.Convert.ToInt32(args[++i]);
if (0 >= maxSplits)
{
maxSplits = int.MaxValue;
}
sampler = new InputSampler.IntervalSampler <K, V>(pcnt, maxSplits);
}
else
{
otherArgs.AddItem(args[i]);
}
}
}
}
}
}
}
catch (FormatException)
{
System.Console.Out.WriteLine("ERROR: Integer expected instead of " + args[i]);
return(PrintUsage());
}
catch (IndexOutOfRangeException)
{
System.Console.Out.WriteLine("ERROR: Required parameter missing from " + args[i -
1]);
return(PrintUsage());
}
}
if (job.GetNumReduceTasks() <= 1)
{
System.Console.Error.WriteLine("Sampler requires more than one reducer");
return(PrintUsage());
}
if (otherArgs.Count < 2)
{
System.Console.Out.WriteLine("ERROR: Wrong number of parameters: ");
return(PrintUsage());
}
if (null == sampler)
{
sampler = new InputSampler.RandomSampler <K, V>(0.1, 10000, 10);
}
Path outf = new Path(otherArgs.Remove(otherArgs.Count - 1));
TotalOrderPartitioner.SetPartitionFile(GetConf(), outf);
foreach (string s in otherArgs)
{
FileInputFormat.AddInputPath(job, new Path(s));
}
InputSampler.WritePartitionFile <K, V>(job, sampler);
return(0);
}
