Examples with FieldSpec com.linkedin.pinot.common.data.FieldSpec used on opensource projects

Example 21 with FieldSpec

use of com.linkedin.pinot.common.data.FieldSpec in project pinot by linkedin.

the class DictionariesTest method testUTF8Characters.

/**
   * Test for ensuring that Strings with special characters can be handled
   * correctly.
   *
   * @throws Exception
   */
@Test
public void testUTF8Characters() throws Exception {
    File indexDir = new File("/tmp/dict.test");
    indexDir.deleteOnExit();
    FieldSpec fieldSpec = new DimensionFieldSpec("test", DataType.STRING, true);
    String[] inputStrings = new String[3];
    char paddingChar = '%';
    // "Café";
    inputStrings[0] = new String(new byte[] { 67, 97, 102, -61, -87 });
    // "François";
    inputStrings[1] = new String(new byte[] { 70, 114, 97, 110, -61, -89, 111, 105, 115 });
    // "Côte d'Ivoire";
    inputStrings[2] = new String(new byte[] { 67, -61, -76, 116, 101, 32, 100, 39, 73, 118, 111, 105, 114, 101 });
    Arrays.sort(inputStrings);
    SegmentDictionaryCreator dictionaryCreator = new SegmentDictionaryCreator(false, inputStrings, fieldSpec, indexDir, paddingChar);
    dictionaryCreator.build(new boolean[] { false });
    for (String inputString : inputStrings) {
        Assert.assertTrue(dictionaryCreator.indexOfSV(inputString) >= 0, "Value not found in dictionary " + inputString);
    }
    dictionaryCreator.close();
    FileUtils.deleteQuietly(indexDir);
}

Example 22 with FieldSpec

use of com.linkedin.pinot.common.data.FieldSpec in project pinot by linkedin.

the class DictionariesTest method testStringColumnPreIndexStatsCollectorForRandomString.

@Test
public void testStringColumnPreIndexStatsCollectorForRandomString() throws Exception {
    FieldSpec spec = new DimensionFieldSpec("column1", DataType.STRING, true);
    AbstractColumnStatisticsCollector statsCollector = new StringColumnPreIndexStatsCollector(spec);
    statsCollector.collect("a");
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect("b");
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect("c");
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect("d");
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect("d");
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect("b");
    Assert.assertFalse(statsCollector.isSorted());
    statsCollector.collect("z");
    Assert.assertFalse(statsCollector.isSorted());
    statsCollector.collect("u");
    Assert.assertFalse(statsCollector.isSorted());
    statsCollector.seal();
    Assert.assertEquals(statsCollector.getCardinality(), 6);
    Assert.assertEquals((statsCollector.getMinValue()).toString(), "a");
    Assert.assertEquals((statsCollector.getMaxValue()).toString(), "z");
    Assert.assertFalse(statsCollector.isSorted());
}

Example 23 with FieldSpec

use of com.linkedin.pinot.common.data.FieldSpec in project pinot by linkedin.

the class DictionariesTest method testLongColumnPreIndexStatsCollector.

@Test
public void testLongColumnPreIndexStatsCollector() throws Exception {
    FieldSpec spec = new DimensionFieldSpec("column1", DataType.LONG, true);
    AbstractColumnStatisticsCollector statsCollector = new LongColumnPreIndexStatsCollector(spec);
    statsCollector.collect(new Integer(1));
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect(new Float(2));
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect(new Long(3));
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect(new Double(4));
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect(new Integer(4));
    Assert.assertTrue(statsCollector.isSorted());
    statsCollector.collect(new Float(2));
    Assert.assertFalse(statsCollector.isSorted());
    statsCollector.collect(new Double(40));
    Assert.assertFalse(statsCollector.isSorted());
    statsCollector.collect(new Double(20));
    Assert.assertFalse(statsCollector.isSorted());
    statsCollector.seal();
    Assert.assertEquals(statsCollector.getCardinality(), 6);
    Assert.assertEquals(((Number) statsCollector.getMinValue()).intValue(), 1);
    Assert.assertEquals(((Number) statsCollector.getMaxValue()).intValue(), 40);
    Assert.assertFalse(statsCollector.isSorted());
}

Example 24 with FieldSpec

use of com.linkedin.pinot.common.data.FieldSpec in project pinot by linkedin.

the class DictionariesTest method testStringsValuesWithPadding.

/**
   * Tests DictionaryCreator for case when one value is a substring of another.
   * For example, in case of sorted values {"abc", "abc def"} after padding,
   * the sorted order would change to {"abc def%%%%", "abc%%%%%%%"}
   *
   * This test asserts that DictionaryCreator.indexOfSV("abc") returns 1 (ie index of "abc%%%%%%%"
   * in actual padded dictionary), and not 0.
   *
   * @throws Exception
   */
@Test
public void testStringsValuesWithPadding() throws Exception {
    File indexDir = new File("/tmp/dict.test");
    indexDir.deleteOnExit();
    FieldSpec fieldSpec = new DimensionFieldSpec("test", DataType.STRING, true);
    String[] inputStrings = new String[2];
    String[] paddedStrings = new String[2];
    char paddingChar = '%';
    inputStrings[0] = "abc def";
    inputStrings[1] = "abc";
    // Sorted order: {"abc", "abc def"}
    Arrays.sort(inputStrings);
    boolean[] isSorted = new boolean[1];
    isSorted[0] = true;
    SegmentDictionaryCreator dictionaryCreator = new SegmentDictionaryCreator(false, inputStrings, fieldSpec, indexDir, paddingChar);
    dictionaryCreator.build(isSorted);
    Assert.assertFalse(isSorted[0]);
    // Get the padded string as stored in the dictionary.
    int targetPaddedLength = dictionaryCreator.getStringColumnMaxLength();
    for (int i = 0; i < inputStrings.length; i++) {
        paddedStrings[i] = SegmentDictionaryCreator.getPaddedString(inputStrings[i], targetPaddedLength, paddingChar);
    }
    // Sorted Order: {"abc def%%%%", "abc%%%%%%%"}
    Arrays.sort(paddedStrings);
    // Assert that indexOfSV for un-padded string returns the index of the corresponding padded string.
    for (int i = 0; i < inputStrings.length; i++) {
        int paddedIndex = dictionaryCreator.indexOfSV(inputStrings[i]);
        Assert.assertTrue(paddedStrings[paddedIndex].equals(SegmentDictionaryCreator.getPaddedString(inputStrings[i], targetPaddedLength, paddingChar)));
    }
    dictionaryCreator.close();
    FileUtils.deleteQuietly(indexDir);
}

Example 25 with FieldSpec

use of com.linkedin.pinot.common.data.FieldSpec in project pinot by linkedin.

the class DictionariesTest method testSingleNullString.

/**
   * Tests SegmentDictionaryCreator for case when there is only one string
   * and it is "null"
   *
   * This test asserts that the padded length of the null string is 4
   *
   * @throws Exception
   */
@Test
public void testSingleNullString() throws Exception {
    File indexDir = new File("/tmp/dict.test");
    FieldSpec fieldSpec = new DimensionFieldSpec("test", DataType.STRING, true);
    String[] inputStrings = new String[1];
    String[] paddedStrings = new String[1];
    inputStrings[0] = "null";
    // Sorted order: {"null"}
    Arrays.sort(inputStrings);
    try {
        SegmentDictionaryCreator dictionaryCreator = new SegmentDictionaryCreator(false, inputStrings, fieldSpec, indexDir, V1Constants.Str.DEFAULT_STRING_PAD_CHAR);
        boolean[] isSorted = new boolean[1];
        isSorted[0] = true;
        dictionaryCreator.build(isSorted);
        // Get the padded string as stored in the dictionary.
        int targetPaddedLength = dictionaryCreator.getStringColumnMaxLength();
        Assert.assertTrue(targetPaddedLength == 4);
        for (int i = 0; i < inputStrings.length; i++) {
            paddedStrings[i] = SegmentDictionaryCreator.getPaddedString(inputStrings[i], targetPaddedLength, V1Constants.Str.DEFAULT_STRING_PAD_CHAR);
        }
        // Sorted Order: {"null"}
        Arrays.sort(paddedStrings);
        // Assert that indexOfSV for un-padded string returns the index of the corresponding padded string.
        for (int i = 0; i < inputStrings.length; i++) {
            int paddedIndex = dictionaryCreator.indexOfSV(inputStrings[i]);
            Assert.assertTrue(paddedStrings[paddedIndex].equals(SegmentDictionaryCreator.getPaddedString(inputStrings[i], targetPaddedLength, V1Constants.Str.DEFAULT_STRING_PAD_CHAR)));
        }
        // Verify that the string "null" did not get changed
        Assert.assertTrue(paddedStrings[0].equals("null"));
        dictionaryCreator.close();
    } catch (Exception e) {
        throw e;
    } finally {
        FileUtils.deleteQuietly(indexDir);
    }
}