[
https://issues.apache.org/jira/browse/MAHOUT-1314?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13825157#comment-13825157
]
Suneel Marthi edited comment on MAHOUT-1314 at 11/18/13 8:14 AM:
-----------------------------------------------------------------
This issue is fixed. Fixing this issue led to the next issue.
The following code in StreamimgKMeansThread.call() is wrong:
{Code}
Iterator<Centroid> datapointsIterator = datapoints.iterator();
if (estimateDistanceCutoff ==
StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF) {
List<Centroid> estimatePoints =
Lists.newArrayListWithExpectedSize(NUM_ESTIMATE_POINTS);
while (datapointsIterator.hasNext() && estimatePoints.size() <
NUM_ESTIMATE_POINTS) {
estimatePoints.add(datapointsIterator.next());
}
estimateDistanceCutoff =
ClusteringUtils.estimateDistanceCutoff(estimatePoints,
searcher.getDistanceMeasure());
}
StreamingKMeans clusterer = new StreamingKMeans(searcher, numClusters,
estimateDistanceCutoff);
while (datapointsIterator.hasNext()) {
clusterer.cluster(datapointsIterator.next());
}
{Code}
On the second loop thru of datapointsIterator, there are no datapoints to
iterate and the job fails with an IllegalArgumentException.
was (Author: smarthi):
This issue is fixed. Fixing this issue led to the next issue.
The following code in StreamimgKMeansThread.call() is wrong:
{Code}
Iterator<Centroid> datapointsIterator = datapoints.iterator();
if (estimateDistanceCutoff ==
StreamingKMeansDriver.INVALID_DISTANCE_CUTOFF) {
List<Centroid> estimatePoints =
Lists.newArrayListWithExpectedSize(NUM_ESTIMATE_POINTS);
while (datapointsIterator.hasNext() && estimatePoints.size() <
NUM_ESTIMATE_POINTS) {
estimatePoints.add(datapointsIterator.next());
}
estimateDistanceCutoff =
ClusteringUtils.estimateDistanceCutoff(estimatePoints,
searcher.getDistanceMeasure());
}
StreamingKMeans clusterer = new StreamingKMeans(searcher, numClusters,
estimateDistanceCutoff);
while (datapointsIterator.hasNext()) {
clusterer.cluster(datapointsIterator.next());
}
{Code}
On the second iteration, there are no datapoints to iterate and the job fails
with an IllegalArgumentException.
> StreamingKMeansReducer throws NullPointerException when
> REDUCE_STREAMING_KMEANS is set to true
> ----------------------------------------------------------------------------------------------
>
> Key: MAHOUT-1314
> URL: https://issues.apache.org/jira/browse/MAHOUT-1314
> Project: Mahout
> Issue Type: Bug
> Components: Clustering
> Affects Versions: 0.8
> Reporter: Adi Haviv
> Assignee: Suneel Marthi
> Fix For: 0.9
>
> Attachments: MAHOUT-1314.patch
>
>
> when REDUCE_STREAMING_KMEANS option is set to true (-rskm) the reducer fails
> with NullPointerException.
> the problem is in the reduce method itself: on line 60 ( return
> input.getCentroid(); )
> it should be input.getCentroid().clone();
> similar to line 81.
> full stack trace:
> java.lang.NullPointerException
> at
> com.google.common.base.Preconditions.checkNotNull(Preconditions.java:191)
> at
> org.apache.mahout.math.random.WeightedThing.<init>(WeightedThing.java:31)
> at
> org.apache.mahout.math.neighborhood.BruteSearch.searchFirst(BruteSearch.java:133)
> at
> org.apache.mahout.clustering.ClusteringUtils.estimateDistanceCutoff(ClusteringUtils.java:100)
> at
> org.apache.mahout.clustering.streaming.mapreduce.StreamingKMeansThread.call(StreamingKMeansThread.java:64)
> at
> org.apache.mahout.clustering.streaming.mapreduce.StreamingKMeansReducer.reduce(StreamingKMeansReducer.java:66)
> at
> org.apache.mahout.clustering.streaming.mapreduce.StreamingKMeansReducer.reduce(StreamingKMeansReducer.java:1)
> at org.apache.hadoop.mapreduce.Reducer.run(Reducer.java:176)
> at
> org.apache.hadoop.mapred.ReduceTask.runNewReducer(ReduceTask.java:650)
> at org.apache.hadoop.mapred.ReduceTask.run(ReduceTask.java:418)
> at
> org.apache.hadoop.mapred.LocalJobRunner$Job.run(LocalJobRunner.java:260)
> it happens every time the REDUCE_STREAMING_KMEANS is set to true.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
