我的输入文件夹有 200 个文件。我希望 MultipleOutputs 将每个文件(使用“map.input.file”标识)的解析输入写入具有相同名称的输出文件中。因为,我没有任何聚合要执行,因此使用 0 reducer 选项(conf .setNumReduceTasks(0))。理想情况下,我应该得到 200 个输出文件。
但是,我的输出有大约 5000 多个文件 - 每个文件只包含一行(流输出)。显然,它不是聚合的。我的假设是,理想情况下,在零减速器中 - 映射器输出应该是聚合的。
帮助表示赞赏。谢谢!
public static void main(String[] args) throws IOException {
if (args.length != 2) {
System.err.println("Usage: MaxTemperature <input path> <output path>");
System.exit(-1);
}
JobConf conf = new JobConf(MultipleOutputEx.class);
conf.setJobName("Duration Count");
FileInputFormat.addInputPath(conf, new Path(args[0]));
FileOutputFormat.setOutputPath(conf, new Path(args[1]));
conf.setNumReduceTasks(0);
conf.setMapperClass(MultipleOutputExMapper.class);
conf.setReducerClass(MultipleOutputExReducer.class);
conf.setMapOutputKeyClass(NullWritable.class);
MultipleOutputs.addMultiNamedOutput(conf,"mofiles", TextOutputFormat.class, NullWritable.class, Text.class);
JobClient.runJob(conf);
}
我的 Mapper 课程是,
public class MultipleOutputExMapper extends MapReduceBase implements
Mapper<LongWritable, Text, NullWritable, Text> {
MultipleOutputs mos = null;
Text fileKey = new Text();
String line = "";
private JobConf conf;
@Override
public void configure(JobConf conf) {
this.conf = conf;
mos = new MultipleOutputs(conf);
}
public void map(LongWritable key, Text value,
OutputCollector<NullWritable, Text> output, Reporter reporter)
throws IOException {
try {
String filename = conf.get("map.input.file");
fileKey.set(filename);
OutputCollector<NullWritable, Text> collector = mos.getCollector(
"mofiles", key.toString(), reporter);
collector.collect(NullWritable.get(), value);
} catch (ArrayIndexOutOfBoundsException E) {
E.printStackTrace();
} catch (Exception E) {
System.out.println(line);
E.printStackTrace();
}
}
@Override
public void close() throws IOException {
mos.close();
}