/** * Sets multiple input paths in the runtime arguments for a file dataset. */ public static void setInputPaths(Map<String, String> arguments, Collection<String> paths) { arguments.remove(INPUT_PATHS); for (String path : paths) { addInputPath(arguments, path); } }
@Override public void initialize() throws Exception { MapReduceContext context = getContext(); String inputPath = context.getRuntimeArguments().get("output.path"); Map<String, String> fileSetArgs = new HashMap<>(); FileSetArguments.addInputPath(fileSetArgs, inputPath); context.addInput(Input.ofDataset(CSV_FILESET_DATASET, fileSetArgs)); context.addOutput(Output.ofDataset(WORDCOUNT_DATASET)); Job job = context.getHadoopJob(); job.setMapperClass(TokenizerMapper.class); job.setReducerClass(IntSumReducer.class); job.setNumReduceTasks(1); } }
FileSetArguments.addInputPath(fileSetArgs, sec.getRuntimeArguments().get("input.path")); JavaPairRDD<LongWritable, Text> input = sec.fromDataset( WorkflowAppWithLocalDatasets.CSV_FILESET_DATASET, fileSetArgs);