mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aleksander Sadecki <aleksander.sade...@pi.esisar.grenoble-inp.fr>
Subject Re: How to list all vectors from a cluster
Date Fri, 23 May 2014 08:36:29 GMT
Hi,

Thank you.

Which version of Apache Mahout you are using? Could you paste here your imports? Thanks

==================================
Projet Industriel PI16 – SICAP
==================================
Equipe: Deschamps Mathias
         Razafindramaka Rado
        Sadecki Aleksander
      

Encadrée par: Brun Emmanuel

Salle C104
==================================
ESISAR
50 rue Barthelemy de Laffemas
BP 54
26902 Valence cedex 9
==================================
tel: 04 56 52 99 16
fax: 04 75 75 94 44
==================================


----- Oryginalna wiadomość -----
Od: "Angel Luis Scull" <ascullp@facinf.uho.edu.cu>
Do: user@mahout.apache.org
Wysłane: czwartek, 22 maj 2014 19:40:24
Temat: Re: How to list all vectors from a cluster

Hi

that work for me
  ...
Path path = new Path(workPath + kmeansClustersPath + 
"/clusteredPoints/part-m-0");
for (Pair<IntWritable, WeightedVectorWritable> record : new 
SequenceFileDirIterable<IntWritable, WeightedVectorWritable>(path, 
PathType.GLOB,
                 PathFilters.logsCRCFilter(), conf)) {
             NamedVector vec = ((NamedVector) 
record.getSecond().getVector());
             System.out.println(record.getFirst().get() + "  " + 
vec.getName());

         }
...


On 22/05/14 11:22, Aleksander Sadecki wrote:
> Hi,
>
> Thank you very much!
>
> I am trying to implement a Java function with this class.
>
> I wrote this piece of code:
>
> 		ClusterDumper dumper = new ClusterDumper(new Path(partMDir), new Path(
> 				seqDir));
>
> 		Map<Integer, List<WeightedPropertyVectorWritable>> dumped = dumper
> 				.getClusterIdToPoints();
>
> 		for (Integer numberOfList : dumped.keySet()) {
> 			List<WeightedPropertyVectorWritable> listWithVectors = dumped
> 					.get(numberOfList);
>
> 			for (WeightedPropertyVectorWritable vec : listWithVectors) {
> 				System.out.println(vec.getVector().toString());
> 			}
> 		}
>
> when I run it, I have got an exception.
>
> Constructor takes 2 parameters:
>
> ClusterDumper(seqFileDir, pointsDir) and I do not know which files should I pass here...
>
> I have got 9 files:
>
> 		String s1 = root + "synthetic_control.data";
> 		String s2 = root + "synthetic_control.seq";
> 		String s3 = root + ".synthetic_control.seq.crc";
> 		String s4 = outputDir + "\\clusteredPoints\\part-m-0";
> 		String s5 = outputDir + "\\clusteredPoints\\.part-m-0.crc";
> 		String s6 = outputDir + "\\clusters-0-final\\_policy";
> 		String s7 = outputDir + "\\clusters-0-final\\part-r-00000";
> 		String s8 = outputDir + "\\clusters-0-final\\._policy.crc";
> 		String s9 = outputDir + "\\clusters-0-final\\.part-r-00000.crc";
>
> 		Path p1 = new Path(s1);
> 		Path p2 = new Path(s2);
> 		Path p3 = new Path(s3);
> 		Path p4 = new Path(s4);
> 		Path p5 = new Path(s5);
> 		Path p6 = new Path(s6);
> 		Path p7 = new Path(s7);
> 		Path p8 = new Path(s8);
> 		Path p9 = new Path(s9);
>
> I tried to find which 2 should I use but nothing works.
>
> Some of them gives me:
>
> synthetic_control.data not a SequenceFile
>
> another one:
>
> org.apache.hadoop.io.LongWritable cannot be cast to org.apache.hadoop.io.IntWritable
>
> or sometimes there is no excpetion but output is empty.
>
> Could you help me?
>
> Thank you in advance


Mime
View raw message