Java Code Examples for org.datavec.api.records.reader.impl.csv.CSVRecordReader#hasNext()

@Test
public void testReset() throws Exception {
    CSVRecordReader rr = new CSVRecordReader(0, ',');
    rr.initialize(new FileSplit(new ClassPathResource("iris.dat").getFile()));

    int nResets = 5;
    for (int i = 0; i < nResets; i++) {

        int lineCount = 0;
        while (rr.hasNext()) {
            List<Writable> line = rr.next();
            assertEquals(5, line.size());
            lineCount++;
        }
        assertFalse(rr.hasNext());
        assertEquals(150, lineCount);
        rr.reset();
    }
}
 

@Test
public void testStreamReset() throws Exception {
    CSVRecordReader rr = new CSVRecordReader(0, ',');
    rr.initialize(new InputStreamInputSplit(new ClassPathResource("iris.dat").getInputStream()));

    int count = 0;
    while(rr.hasNext()){
        assertNotNull(rr.next());
        count++;
    }
    assertEquals(150, count);

    assertFalse(rr.resetSupported());

    try{
        rr.reset();
        fail("Expected exception");
    } catch (Exception e){
        e.printStackTrace();
    }
}
 

@Test
public void testRegex() throws Exception {
    CSVRecordReader reader = new CSVRegexRecordReader(0, ",", null, new String[] {null, "(.+) (.+) (.+)"});
    reader.initialize(new StringSplit("normal,1.2.3.4 space separator"));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        assertEquals("Entry count", 4, vals.size());
        assertEquals("normal", vals.get(0).toString());
        assertEquals("1.2.3.4", vals.get(1).toString());
        assertEquals("space", vals.get(2).toString());
        assertEquals("separator", vals.get(3).toString());
    }
}
 

@Test
public void testTabsAsSplit1() throws Exception {

    CSVRecordReader reader = new CSVRecordReader(0, '\t');
    reader.initialize(new FileSplit(new ClassPathResource("datavec-api/tabbed.txt").getFile()));
    while (reader.hasNext()) {
        List<Writable> list = new ArrayList<>(reader.next());

        assertEquals(2, list.size());
    }
}
 

@Test
public void testRegex() throws Exception {
    CSVRecordReader reader = new CSVRegexRecordReader(0, ",", null, new String[] {null, "(.+) (.+) (.+)"});
    reader.initialize(new StringSplit("normal,1.2.3.4 space separator"));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        assertEquals("Entry count", 4, vals.size());
        assertEquals("normal", vals.get(0).toString());
        assertEquals("1.2.3.4", vals.get(1).toString());
        assertEquals("space", vals.get(2).toString());
        assertEquals("separator", vals.get(3).toString());
    }
}
 

@Test
public void testPipesAsSplit() throws Exception {

    CSVRecordReader reader = new CSVRecordReader(0, '|');
    reader.initialize(new FileSplit(new ClassPathResource("datavec-api/issue414.csv").getFile()));
    int lineidx = 0;
    List<Integer> sixthColumn = Arrays.asList(13, 95, 15, 25);
    while (reader.hasNext()) {
        List<Writable> list = new ArrayList<>(reader.next());

        assertEquals(10, list.size());
        assertEquals((long)sixthColumn.get(lineidx), list.get(5).toInt());
        lineidx++;
    }
}
 

@Test
public void testEmptyEntries() throws Exception {
    CSVRecordReader reader = new CSVRecordReader();
    reader.initialize(new StringSplit("1,1,8.0,,,,14.0,,,,15.0,,,,,,,,,,,,"));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        assertEquals("Entry count", 23, vals.size());
    }
}
 

@Test
public void testNext() throws Exception {
    CSVRecordReader reader = new CSVRecordReader();
    reader.initialize(new StringSplit("1,1,8.0,,,,14.0,,,,15.0,,,,,,,,,,,,1"));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        List<Writable> arr = new ArrayList<>(vals);

        assertEquals("Entry count", 23, vals.size());
        Text lastEntry = (Text) arr.get(arr.size() - 1);
        assertEquals("Last entry garbage", 1, lastEntry.getLength());
    }
}
 

@Test
public void testWrite() throws Exception {
    File tempFile = File.createTempFile("datavec", "writer");
    tempFile.deleteOnExit();
    FileSplit fileSplit = new FileSplit(tempFile);
    CSVRecordWriter writer = new CSVRecordWriter();
    writer.initialize(fileSplit,new NumberOfRecordsPartitioner());
    List<Writable> collection = new ArrayList<>();
    collection.add(new Text("12"));
    collection.add(new Text("13"));
    collection.add(new Text("14"));

    writer.write(collection);

    CSVRecordReader reader = new CSVRecordReader(0);
    reader.initialize(new FileSplit(tempFile));
    int cnt = 0;
    while (reader.hasNext()) {
        List<Writable> line = new ArrayList<>(reader.next());
        assertEquals(3, line.size());

        assertEquals(12, line.get(0).toInt());
        assertEquals(13, line.get(1).toInt());
        assertEquals(14, line.get(2).toInt());
        cnt++;
    }
    assertEquals(1, cnt);
}
 

@Test
public void testWithQuotes() throws Exception {
    CSVRecordReader reader = new CSVRecordReader(0, ',', '\"');
    reader.initialize(new StringSplit("1,0,3,\"Braund, Mr. Owen Harris\",male,\"\"\"\""));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        assertEquals("Entry count", 6, vals.size());
        assertEquals("1", vals.get(0).toString());
        assertEquals("0", vals.get(1).toString());
        assertEquals("3", vals.get(2).toString());
        assertEquals("Braund, Mr. Owen Harris", vals.get(3).toString());
        assertEquals("male", vals.get(4).toString());
        assertEquals("\"", vals.get(5).toString());
    }
}
 

@Test
public void testLineRecordReader() throws Exception {

    File dataFile = new ClassPathResource("iris.dat").getFile();
    List<String> lines = FileUtils.readLines(dataFile);

    JavaSparkContext sc = getContext();
    JavaRDD<String> linesRdd = sc.parallelize(lines);

    CSVRecordReader rr = new CSVRecordReader(0, ',');

    JavaRDD<List<Writable>> out = linesRdd.map(new LineRecordReaderFunction(rr));
    List<List<Writable>> outList = out.collect();


    CSVRecordReader rr2 = new CSVRecordReader(0, ',');
    rr2.initialize(new FileSplit(dataFile));
    Set<List<Writable>> expectedSet = new HashSet<>();
    int totalCount = 0;
    while (rr2.hasNext()) {
        expectedSet.add(rr2.next());
        totalCount++;
    }

    assertEquals(totalCount, outList.size());

    for (List<Writable> line : outList) {
        assertTrue(expectedSet.contains(line));
    }
}
 

@Test
public void testLineRecordReader() throws Exception {

    File dataFile = new ClassPathResource("iris.dat").getFile();
    List<String> lines = FileUtils.readLines(dataFile);

    List<String> linesRdd = (lines);

    CSVRecordReader rr = new CSVRecordReader(0, ',');

    List<List<Writable>> out = linesRdd.stream().map(input -> new LineRecordReaderFunction(rr).apply(input)).collect(Collectors.toList());
    List<List<Writable>> outList = out;


    CSVRecordReader rr2 = new CSVRecordReader(0, ',');
    rr2.initialize(new FileSplit(dataFile));
    Set<List<Writable>> expectedSet = new HashSet<>();
    int totalCount = 0;
    while (rr2.hasNext()) {
        expectedSet.add(rr2.next());
        totalCount++;
    }

    assertEquals(totalCount, outList.size());

    for (List<Writable> line : outList) {
        assertTrue(expectedSet.contains(line));
    }
}
 

@Test
public void testCsvSimple() throws Exception {
    File dir = testDir.newFolder();
    File f1 = new File(dir, "file1.txt");
    File f2 = new File(dir, "file2.txt");

    FileUtils.writeStringToFile(f1, "a,b,c\nd,e,f", StandardCharsets.UTF_8);
    FileUtils.writeStringToFile(f2, "1,2,3", StandardCharsets.UTF_8);

    List<URI> uris = Arrays.asList(f1.toURI(), f2.toURI());

    CSVRecordReader rr = new CSVRecordReader();

    TestStreamFunction fn = new TestStreamFunction();
    InputSplit is = new StreamInputSplit(uris, fn);
    rr.initialize(is);

    List<List<Writable>> exp = new ArrayList<>();
    exp.add(Arrays.<Writable>asList(new Text("a"), new Text("b"), new Text("c")));
    exp.add(Arrays.<Writable>asList(new Text("d"), new Text("e"), new Text("f")));
    exp.add(Arrays.<Writable>asList(new Text("1"), new Text("2"), new Text("3")));

    List<List<Writable>> act = new ArrayList<>();
    while(rr.hasNext()){
        act.add(rr.next());
    }

    assertEquals(exp, act);

    //Check that the specified stream loading function was used, not the default:
    assertEquals(uris, fn.calledWithUris);

    rr.reset();
    int count = 0;
    while(rr.hasNext()) {
        count++;
        rr.next();
    }
    assertEquals(3, count);
}
 

@Test
public void testTabsAsSplit1() throws Exception {

    CSVRecordReader reader = new CSVRecordReader(0, '\t');
    reader.initialize(new FileSplit(new ClassPathResource("/tabbed.txt").getFile()));
    while (reader.hasNext()) {
        List<Writable> list = new ArrayList<>(reader.next());

        assertEquals(2, list.size());
    }
}
 

@Test
public void testLineRecordReader() throws Exception {

    File dataFile = new ClassPathResource("iris.dat").getFile();
    List<String> lines = FileUtils.readLines(dataFile);

    JavaSparkContext sc = getContext();
    JavaRDD<String> linesRdd = sc.parallelize(lines);

    CSVRecordReader rr = new CSVRecordReader(0, ',');

    JavaRDD<List<Writable>> out = linesRdd.map(new LineRecordReaderFunction(rr));
    List<List<Writable>> outList = out.collect();


    CSVRecordReader rr2 = new CSVRecordReader(0, ',');
    rr2.initialize(new FileSplit(dataFile));
    Set<List<Writable>> expectedSet = new HashSet<>();
    int totalCount = 0;
    while (rr2.hasNext()) {
        expectedSet.add(rr2.next());
        totalCount++;
    }

    assertEquals(totalCount, outList.size());

    for (List<Writable> line : outList) {
        assertTrue(expectedSet.contains(line));
    }
}
 

@Test
public void testEmptyEntries() throws Exception {
    CSVRecordReader reader = new CSVRecordReader();
    reader.initialize(new StringSplit("1,1,8.0,,,,14.0,,,,15.0,,,,,,,,,,,,"));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        assertEquals("Entry count", 23, vals.size());
    }
}
 

@Test
public void testNext() throws Exception {
    CSVRecordReader reader = new CSVRecordReader();
    reader.initialize(new StringSplit("1,1,8.0,,,,14.0,,,,15.0,,,,,,,,,,,,1"));
    while (reader.hasNext()) {
        List<Writable> vals = reader.next();
        List<Writable> arr = new ArrayList<>(vals);

        assertEquals("Entry count", 23, vals.size());
        Text lastEntry = (Text) arr.get(arr.size() - 1);
        assertEquals("Last entry garbage", 1, lastEntry.getLength());
    }
}
 

@Test
public void testWrite() throws Exception {
    File tempFile = File.createTempFile("datavec", "writer");
    tempFile.deleteOnExit();
    FileSplit fileSplit = new FileSplit(tempFile);
    CSVRecordWriter writer = new CSVRecordWriter();
    writer.initialize(fileSplit,new NumberOfRecordsPartitioner());
    List<Writable> collection = new ArrayList<>();
    collection.add(new Text("12"));
    collection.add(new Text("13"));
    collection.add(new Text("14"));

    writer.write(collection);

    CSVRecordReader reader = new CSVRecordReader(0);
    reader.initialize(new FileSplit(tempFile));
    int cnt = 0;
    while (reader.hasNext()) {
        List<Writable> line = new ArrayList<>(reader.next());
        assertEquals(3, line.size());

        assertEquals(12, line.get(0).toInt());
        assertEquals(13, line.get(1).toInt());
        assertEquals(14, line.get(2).toInt());
        cnt++;
    }
    assertEquals(1, cnt);
}
 

@Test
public void testMeta() throws Exception {
    CSVRecordReader rr = new CSVRecordReader(0, ',');
    rr.initialize(new FileSplit(new ClassPathResource("iris.dat").getFile()));

    int lineCount = 0;
    List<RecordMetaData> metaList = new ArrayList<>();
    List<List<Writable>> writables = new ArrayList<>();
    while (rr.hasNext()) {
        Record r = rr.nextRecord();
        assertEquals(5, r.getRecord().size());
        lineCount++;
        RecordMetaData meta = r.getMetaData();
        //            System.out.println(r.getRecord() + "\t" + meta.getLocation() + "\t" + meta.getURI());

        metaList.add(meta);
        writables.add(r.getRecord());
    }
    assertFalse(rr.hasNext());
    assertEquals(150, lineCount);
    rr.reset();


    System.out.println("\n\n\n--------------------------------");
    List<Record> contents = rr.loadFromMetaData(metaList);
    assertEquals(150, contents.size());
    //        for(Record r : contents ){
    //            System.out.println(r);
    //        }

    List<RecordMetaData> meta2 = new ArrayList<>();
    meta2.add(metaList.get(100));
    meta2.add(metaList.get(90));
    meta2.add(metaList.get(80));
    meta2.add(metaList.get(70));
    meta2.add(metaList.get(60));

    List<Record> contents2 = rr.loadFromMetaData(meta2);
    assertEquals(writables.get(100), contents2.get(0).getRecord());
    assertEquals(writables.get(90), contents2.get(1).getRecord());
    assertEquals(writables.get(80), contents2.get(2).getRecord());
    assertEquals(writables.get(70), contents2.get(3).getRecord());
    assertEquals(writables.get(60), contents2.get(4).getRecord());
}
 

@Test
public void testMeta() throws Exception {
    CSVRecordReader rr = new CSVRecordReader(0, ',');
    rr.initialize(new FileSplit(new ClassPathResource("datavec-api/iris.dat").getFile()));

    int lineCount = 0;
    List<RecordMetaData> metaList = new ArrayList<>();
    List<List<Writable>> writables = new ArrayList<>();
    while (rr.hasNext()) {
        Record r = rr.nextRecord();
        assertEquals(5, r.getRecord().size());
        lineCount++;
        RecordMetaData meta = r.getMetaData();
        //            System.out.println(r.getRecord() + "\t" + meta.getLocation() + "\t" + meta.getURI());

        metaList.add(meta);
        writables.add(r.getRecord());
    }
    assertFalse(rr.hasNext());
    assertEquals(150, lineCount);
    rr.reset();


    System.out.println("\n\n\n--------------------------------");
    List<Record> contents = rr.loadFromMetaData(metaList);
    assertEquals(150, contents.size());
    //        for(Record r : contents ){
    //            System.out.println(r);
    //        }

    List<RecordMetaData> meta2 = new ArrayList<>();
    meta2.add(metaList.get(100));
    meta2.add(metaList.get(90));
    meta2.add(metaList.get(80));
    meta2.add(metaList.get(70));
    meta2.add(metaList.get(60));

    List<Record> contents2 = rr.loadFromMetaData(meta2);
    assertEquals(writables.get(100), contents2.get(0).getRecord());
    assertEquals(writables.get(90), contents2.get(1).getRecord());
    assertEquals(writables.get(80), contents2.get(2).getRecord());
    assertEquals(writables.get(70), contents2.get(3).getRecord());
    assertEquals(writables.get(60), contents2.get(4).getRecord());
}