Java 类org.apache.hadoop.mapred.lib.InputSampler 实例源码

Object[] getPivots(Path input, int numPivs) throws IOException
{

    JobConf job = new JobConf();
    job.setInputFormat(KeyValueTextInputFormat.class);
    job.setMapOutputKeyClass(Text.class);
    org.apache.hadoop.mapred.FileInputFormat.addInputPath(job, input);
    final KeyValueTextInputFormat inf = (KeyValueTextInputFormat) job.getInputFormat();
    InputSampler.Sampler<Text, Text> sampler = new InputSampler.RandomSampler<Text, Text>(1.0, numPivs, 100);
    Object[] samples = sampler.getSample(inf, job);
    return samples;
}