Package org.apache.crunch.impl.mr

Examples of org.apache.crunch.impl.mr.MRPipeline

55
56
57
58
59
60
61
62
    MemPipeline.clearCounters();
  }
 
  @Test
  public void testStageResultsCountersMRWritables() throws Exception {
    testSpecialKeywordCount(new MRPipeline(StageResultsCountersIT.class, tmpDir.getDefaultConfiguration()),
        WritableTypeFamily.getInstance());
  }
View Full Code Here
61
62
63
64
65
66
67
68
        WritableTypeFamily.getInstance());
  }

  @Test
  public void testStageResultsCountersMRAvro() throws Exception {
    testSpecialKeywordCount(new MRPipeline(StageResultsCountersIT.class, tmpDir.getDefaultConfiguration()),
        AvroTypeFamily.getInstance());
  }
View Full Code Here
40
41
42
43
44
45
46
47
48
49
50
51
52
  @Test
  public void testVanillaCSV() throws Exception {
    String[] expectedFileContents = { "1,2,3,4", "5,6,7,8", "9,10,11", "12,13,14" };

    String vanillaCSVFile = tmpDir.copyResourceFileName("vanilla.csv");
    Pipeline pipeline = new MRPipeline(CSVFileSourceIT.class, tmpDir.getDefaultConfiguration());
    PCollection<String> csvLines = pipeline.read(new CSVFileSource(new Path(vanillaCSVFile)));
    pipeline.run();

    Collection<String> csvLinesList = csvLines.asCollection().getValue();

    for (int i = 0; i < expectedFileContents.length; i++) {
      assertTrue(csvLinesList.contains(expectedFileContents[i]));
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
69
70
    String[] expectedFileContents = {
        "\"Champion, Mac\",\"1234 Hoth St.\n\tApartment 101\n\tAtlanta, GA\n\t64086\",\"30\",\"M\",\"5/28/2010 12:00:00 AM\",\"Just some guy\"",
        "\"Champion, Mac\",\"5678 Tatooine Rd. Apt 5, Mobile, AL 36608\",\"30\",\"M\",\"Some other date\",\"short description\"" };

    String csvWithNewlines = tmpDir.copyResourceFileName("withNewlines.csv");
    Pipeline pipeline = new MRPipeline(CSVFileSourceIT.class, tmpDir.getDefaultConfiguration());
    PCollection<String> csvLines = pipeline.read(new CSVFileSource(new Path(csvWithNewlines)));
    pipeline.run();

    Collection<String> csvLinesList = csvLines.asCollection().getValue();

    for (int i = 0; i < expectedFileContents.length; i++) {
      assertTrue(csvLinesList.contains(expectedFileContents[i]));
View Full Code Here
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
    String[] expectedFileContents = {
        "*Champion, Mac*,*1234 Hoth St.\n\tApartment 101\n\tAtlanta, GA\n\t64086*,*30*,*M*,*5/28/2010 12:00:00 AM*,*Just some guy*",
        "*Mac, Champion*,*5678 Tatooine Rd. Apt 5, Mobile, AL 36608*,*30*,*M*,*Some other date*,*short description*" };

    String csvWithNewlines = tmpDir.copyResourceFileName("customQuoteCharWithNewlines.csv");
    Pipeline pipeline = new MRPipeline(CSVFileSourceIT.class, tmpDir.getDefaultConfiguration());
    PCollection<String> csvLines = pipeline.read(new CSVFileSource(new Path(csvWithNewlines),
        CSVLineReader.DEFAULT_BUFFER_SIZE, CSVLineReader.DEFAULT_INPUT_FILE_ENCODING, '*', '*',
        CSVLineReader.DEFAULT_ESCAPE_CHARACTER));
    pipeline.run();

    Collection<String> csvLinesList = csvLines.asCollection().getValue();

    for (int i = 0; i < expectedFileContents.length; i++) {
      assertTrue(csvLinesList.contains(expectedFileContents[i]));
View Full Code Here
105
106
107
108
109
110
111
112
113
114
115
116
117
118
  public void testBrokenLineParsingInChinese() throws IOException {
    final String[] expectedChineseLines = { "您好我叫马克，我从亚拉巴马州来，我是软件工程师，我二十八岁", "我有一个宠物，它是一个小猫，它六岁，它很漂亮",
        "我喜欢吃饭，“我觉得这个饭最好\n＊蛋糕\n＊包子\n＊冰淇淋\n＊啤酒“，他们都很好，我也很喜欢奶酪但它是不健康的", "我是男的，我的头发很短，我穿蓝色的裤子，“我穿黑色的、“衣服”" };
    String chineseLines = tmpDir.copyResourceFileName("brokenChineseLines.csv");

    Pipeline pipeline = new MRPipeline(CSVFileSourceIT.class, tmpDir.getDefaultConfiguration());
    PCollection<String> csvLines = pipeline.read(new CSVFileSource(new Path(chineseLines),
        CSVLineReader.DEFAULT_BUFFER_SIZE, CSVLineReader.DEFAULT_INPUT_FILE_ENCODING, '“', '”', '、'));
    pipeline.run();
    Collection<String> csvLinesList = csvLines.asCollection().getValue();
    for (int i = 0; i < expectedChineseLines.length; i++) {
      assertTrue(csvLinesList.contains(expectedChineseLines[i]));
    }
  }
View Full Code Here
98
99
100
101
102
103
104
105
106
107
108
    runMapsideLeftOuterJoin(MemPipeline.getInstance(), true);
  }

  @Test
  public void testMapsideJoin_RightSideIsEmpty() throws IOException {
    MRPipeline pipeline = new MRPipeline(MapsideJoinStrategyIT.class, tmpDir.getDefaultConfiguration());
    PTable<Integer, String> customerTable = readTable(pipeline, "customers.txt");
    PTable<Integer, String> orderTable = readTable(pipeline, "orders.txt");

    PTable<Integer, String> filteredOrderTable = orderTable
        .parallelDo(FilterFns.<Pair<Integer, String>>REJECT_ALL(), orderTable.getPTableType());
View Full Code Here
116
117
118
119
120
121
122
    assertTrue(materializedJoin.isEmpty());
  }

  @Test
  public void testMapsideJoin() throws IOException {
    runMapsideJoin(new MRPipeline(MapsideJoinStrategyIT.class, tmpDir.getDefaultConfiguration()), false);
  }
View Full Code Here
121
122
123
124
125
126
127
    runMapsideJoin(new MRPipeline(MapsideJoinStrategyIT.class, tmpDir.getDefaultConfiguration()), false);
  }
 
  @Test
  public void testMapsideJoin_LeftOuterJoin() throws IOException {
    runMapsideLeftOuterJoin(new MRPipeline(MapsideJoinStrategyIT.class, tmpDir.getDefaultConfiguration()), false);
  }
View Full Code Here
91
92
93
94
95
96
97
98
    return new UnionCollection<S>(internal);
  }

  @Override
  public <T> PCollection<T> parallelDo(DoFn<S, T> fn, PType<T> type) {
    MRPipeline pipeline = (MRPipeline) getPipeline();
    return parallelDo("S" + pipeline.getNextAnonymousStageId(), fn, type);
  }
View Full Code Here
TOP

Related Classes of org.apache.crunch.impl.mr.MRPipeline

  • com.cloudera.cdk.data.crunch.TestCrunchDatasets
  • org.apache.bigtop.bigpetstore.etl.CrunchETL
  • org.apache.crunch.Breakpoint2IT
  • org.apache.crunch.BreakpointIT
  • org.apache.crunch.CancelJobsIT
  • org.apache.crunch.CheckpointIT
  • org.apache.crunch.CleanTextIT
  • org.apache.crunch.CollectionPObjectIT
  • org.apache.crunch.CollectionsIT
  • org.apache.crunch.CollectionsLengthIT
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.