Getting java.lang.IllegalArgumentException: requirement failed while calling Sparks MLLIB StreamingKMeans from java application

SeanB picture SeanB · Jun 9, 2015 · Viewed 8k times · Source

I am new to Spark and MLlib and I'm trying to call StreamingKMeans from my java application and I get an exception that I don't seem to understand. Here is my code for transforming my training data:

JavaDStream<Vector> trainingData = sjsc.textFileStream("/training")
            .map(new Function<String, Vector>() {
                public DenseVector call(String line) throws Exception {
                    String[] lineSplit = line.split(",");

                    double[] doubleValues = new double[lineSplit.length];
                    for (int i = 0; i < lineSplit.length; i++) {
                        doubleValues[i] = Double.parseDouble(lineSplit[i] != null ? !""
                                .equals(lineSplit[i]) ? lineSplit[i] : "0" : "0");
                    }
                    DenseVector denseV = new DenseVector(doubleValues);
                    if (denseV.size() != 16) {
                        throw new Exception("All vectors are not the same size!");
                    }
                    System.out.println("Vector length is:" + denseV.size());
                    return denseV;
                }
            });

Here the code where I call the trainOn method:

int numDimensions = 18;
int numClusters = 2;
StreamingKMeans model = new StreamingKMeans();
model.setK(numClusters);
model.setDecayFactor(.5);
model.setRandomCenters(numDimensions, 0.0, Utils.random().nextLong());

model.trainOn(trainingData.dstream());

And here is the exception I receive:

java.lang.IllegalArgumentException: requirement failed
    at scala.Predef$.require(Predef.scala:221)
    at org.apache.spark.mllib.util.MLUtils$.fastSquaredDistance(MLUtils.scala:292)
    at org.apache.spark.mllib.clustering.KMeans$.fastSquaredDistance(KMeans.scala:485)
    at org.apache.spark.mllib.clustering.KMeans$$anonfun$findClosest$1.apply(KMeans.scala:459)
    at org.apache.spark.mllib.clustering.KMeans$$anonfun$findClosest$1.apply(KMeans.scala:453)
    at scala.collection.mutable.ArraySeq.foreach(ArraySeq.scala:73)
    at org.apache.spark.mllib.clustering.KMeans$.findClosest(KMeans.scala:453)
    at org.apache.spark.mllib.clustering.KMeansModel.predict(KMeansModel.scala:35)
    at org.apache.spark.mllib.clustering.StreamingKMeans$$anonfun$predictOnValues$1.apply(StreamingKMeans.scala:258)
    at org.apache.spark.mllib.clustering.StreamingKMeans$$anonfun$predictOnValues$1.apply(StreamingKMeans.scala:258)
    at org.apache.spark.rdd.PairRDDFunctions$$anonfun$mapValues$1$$anonfun$apply$15.apply(PairRDDFunctions.scala:674)
    at org.apache.spark.rdd.PairRDDFunctions$$anonfun$mapValues$1$$anonfun$apply$15.apply(PairRDDFunctions.scala:674)
    at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
    at scala.collection.Iterator$$anon$10.next(Iterator.scala:312)
    at scala.collection.Iterator$class.foreach(Iterator.scala:727)
    at scala.collection.AbstractIterator.foreach(Iterator.scala:1157)
    at scala.collection.generic.Growable$class.$plus$plus$eq(Growable.scala:48)
    at scala.collection.mutable.ArrayBuffer.$plus$plus$eq(ArrayBuffer.scala:103)
    at scala.collection.mutable.ArrayBuffer.$plus$plus$eq(ArrayBuffer.scala:47)
    at scala.collection.TraversableOnce$class.to(TraversableOnce.scala:273)
    at scala.collection.AbstractIterator.to(Iterator.scala:1157)
    at scala.collection.TraversableOnce$class.toBuffer(TraversableOnce.scala:265)
    at scala.collection.AbstractIterator.toBuffer(Iterator.scala:1157)
    at scala.collection.TraversableOnce$class.toArray(TraversableOnce.scala:252)
    at scala.collection.AbstractIterator.toArray(Iterator.scala:1157)
    at org.apache.spark.rdd.RDD$$anonfun$33.apply(RDD.scala:1177)
    at org.apache.spark.rdd.RDD$$anonfun$33.apply(RDD.scala:1177)
    at org.apache.spark.SparkContext$$anonfun$runJob$5.apply(SparkContext.scala:1498)
    at org.apache.spark.SparkContext$$anonfun$runJob$5.apply(SparkContext.scala:1498)
    at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:61)
    at org.apache.spark.scheduler.Task.run(Task.scala:64)
    at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:203)
    at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:895)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:918)
    at java.lang.Thread.run(Thread.java:662)

As you can see in the above code I am checking to make sure my vectors are all the same size and they appear to be, even though the error is suggesting they are not. Any help would be greatly appreciated!

Answer

vito picture vito · Jun 5, 2016

All the vectors are not of the same dimension could cause this exception.

In my experience, another possible reason is Vector which contains the value of NaN.

All values in the vector can not contain NaN.