List of usage examples for org.apache.hadoop.mapreduce.lib.output MultipleOutputs write
@SuppressWarnings("unchecked") public void write(KEYOUT key, VALUEOUT value, String baseOutputPath) throws IOException, InterruptedException
From source file:be.uantwerpen.adrem.bigfim.AprioriPhaseReducerTest.java
License:Apache License
private void addToOutput(MultipleOutputs<Text, Text> mos, String prefix, int support) throws IOException, InterruptedException { mos.write(new Text(prefix), new Text("" + support), "base/tg0/trieGroup-0"); mos.write(new Text(prefix), new Text("" + support), "fis"); }
From source file:be.uantwerpen.adrem.bigfim.ComputeTidListReducerTest.java
License:Apache License
@Test public void One_PG_One_Item() throws Exception { MultipleOutputs<IntArrayWritable, IntMatrixWritable> mos = createMock(MultipleOutputs.class); mos.write(newIAW(1), EmptyImw, "pg/bucket-0"); mos.write(newIAW(0), new IntMatrixWritable(newIAW(0, 1, 2, 4, 7, 9), newIAW(0, 1, 2, 3, 5, 6, 8)), "pg/bucket-0"); mos.write(EmptyIaw, EmptyImw, "pg/bucket-0"); mos.close();/* ww w. j a v a 2s .c om*/ Reducer.Context ctx = createMock(Reducer.Context.class); EasyMock.expect(ctx.getConfiguration()).andReturn(createConfiguration()).anyTimes(); EasyMock.expect(ctx.getTaskAttemptID()).andReturn(new TaskAttemptID()).anyTimes(); EasyMock.replay(ctx, mos); ComputeTidListReducer reducer = new ComputeTidListReducer(); reducer.setup(ctx); setField(reducer, "mos", mos); reducer.reduce(new Text("1"), createTestInput_1Item(), ctx); reducer.cleanup(ctx); EasyMock.verify(mos); }
From source file:be.uantwerpen.adrem.bigfim.ComputeTidListReducerTest.java
License:Apache License
@Test public void One_PG_N_Items() throws Exception { MultipleOutputs<IntArrayWritable, IntMatrixWritable> mos = createMock(MultipleOutputs.class); mos.write(newIAW(1), EmptyImw, "pg/bucket-0"); mos.write(newIAW(0), new IntMatrixWritable(newIAW(0, 1, 2, 4, 7, 9), newIAW(0, 1, 2, 3, 5, 6, 8)), "pg/bucket-0"); mos.write(newIAW(1), new IntMatrixWritable(newIAW(1, 2, 3), newIAW(4, 5, 6)), "pg/bucket-0"); mos.write(newIAW(3), new IntMatrixWritable(newIAW(4, 7, 9), newIAW(4, 7, 9)), "pg/bucket-0"); mos.write(EmptyIaw, EmptyImw, "pg/bucket-0"); mos.close();//from w w w. j a v a 2 s. c o m Reducer.Context ctx = createMock(Reducer.Context.class); EasyMock.expect(ctx.getConfiguration()).andReturn(createConfiguration()).anyTimes(); EasyMock.expect(ctx.getTaskAttemptID()).andReturn(new TaskAttemptID()).anyTimes(); EasyMock.replay(ctx, mos); ComputeTidListReducer reducer = new ComputeTidListReducer(); reducer.setup(ctx); setField(reducer, "mos", mos); reducer.reduce(new Text("1"), createTestInput_NItems(), ctx); reducer.cleanup(ctx); EasyMock.verify(mos); }
From source file:be.uantwerpen.adrem.bigfim.ComputeTidListReducerTest.java
License:Apache License
@Test public void N_PG_N_Items() throws Exception { MultipleOutputs<IntArrayWritable, IntMatrixWritable> mos = createMock(MultipleOutputs.class); mos.write(newIAW(1), EmptyImw, "pg/bucket-0"); mos.write(newIAW(0), new IntMatrixWritable(newIAW(0, 1, 2, 4, 7, 9), newIAW(0, 1, 2, 3, 5, 6, 8)), "pg/bucket-0"); mos.write(newIAW(1), new IntMatrixWritable(newIAW(1, 2, 3), newIAW(4, 5, 6)), "pg/bucket-0"); mos.write(newIAW(3), new IntMatrixWritable(newIAW(4, 7, 9), newIAW(4, 7, 9)), "pg/bucket-0"); mos.write(EmptyIaw, EmptyImw, "pg/bucket-0"); mos.write(newIAW(2), EmptyImw, "pg/bucket-1"); mos.write(newIAW(1), new IntMatrixWritable(newIAW(1, 4, 7, 8), newIAW(1, 5, 6, 8)), "pg/bucket-1"); mos.write(newIAW(2), new IntMatrixWritable(newIAW(3, 5, 7), newIAW(1, 2, 3, 4, 5, 6, 7, 8, 9)), "pg/bucket-1"); mos.write(EmptyIaw, EmptyImw, "pg/bucket-1"); mos.close();// w ww .ja v a 2 s . c o m Reducer.Context ctx = createMock(Reducer.Context.class); EasyMock.expect(ctx.getConfiguration()).andReturn(createConfiguration()).anyTimes(); EasyMock.expect(ctx.getTaskAttemptID()).andReturn(new TaskAttemptID()).anyTimes(); EasyMock.replay(ctx, mos); ComputeTidListReducer reducer = new ComputeTidListReducer(); reducer.setup(ctx); setField(reducer, "mos", mos); reducer.reduce(new Text("1"), createTestInput_NItems(), ctx); reducer.reduce(new Text("2"), createTestInput_NItems2(), ctx); reducer.cleanup(ctx); EasyMock.verify(mos); }
From source file:de.tudarmstadt.ukp.dkpro.c4corpus.hadoop.full.WARCWriterReducerClass.java
License:Apache License
/** * Writes single WARCWritable to the output with specific output file prefix * * @param warcWritable warc record//from ww w . j a v a2 s.co m * @param multipleOutputs output * @throws IOException exception * @throws InterruptedException exception */ // TODO move somewhere else? public static void writeSingleWARCWritableToOutput(WARCWritable warcWritable, MultipleOutputs<NullWritable, WARCWritable> multipleOutputs) throws IOException, InterruptedException { WARCRecord.Header header = warcWritable.getRecord().getHeader(); String license = header.getField(WARCRecord.WARCRecordFieldConstants.LICENSE); String language = header.getField(WARCRecord.WARCRecordFieldConstants.LANGUAGE); String noBoilerplate = header.getField(WARCRecord.WARCRecordFieldConstants.NO_BOILERPLATE); String minimalHtml = header.getField(WARCRecord.WARCRecordFieldConstants.MINIMAL_HTML); // set the file name prefix String fileName = createOutputFilePrefix(license, language, noBoilerplate, minimalHtml); // bottleneck of single reducer for all "Lic_none_Lang_en" pages (majority of Web) // if ("en".equals(language) && LicenseDetector.NO_LICENCE.equals(license)) { // long simHash = Long // .valueOf(header.getField(WARCRecord.WARCRecordFieldConstants.SIMHASH)); // int binNumber = getBinNumberFromSimHash(simHash); // fileName = createOutputFilePrefix(license, language, noBoilerplate); // } multipleOutputs.write(NullWritable.get(), warcWritable, fileName); }
From source file:eu.scape_project.up2ti.output.SimpleKeyValueOutputWriter.java
License:Apache License
/** * Record method for hadoop job./*from w w w . j a va 2 s.co m*/ * * @param resultMap Result map where K: recordkey-identificationtype, V: * tool identificationtype identificationresult) * @param mos Multiple output writer */ @Override public void write(HashMap<String, List<String>> resultMap, MultipleOutputs mos) { Iterator iter = resultMap.keySet().iterator(); while (iter.hasNext()) { String key = (String) iter.next(); List<String> valueList = resultMap.get(key); try { for (String value : valueList) { mos.write("idtab", new Text(key), new Text(value)); } } catch (IOException ex) { LOG.error("I/O Error", ex); } catch (InterruptedException ex) { LOG.error("Interrupted Error", ex); } } }