Examples with DecimalColumnVector org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector used on opensource projects

Example 81 with DecimalColumnVector

use of org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector in project hive by apache.

the class FuncRoundWithNumDigitsDecimalToDecimal method evaluate.

@Override
public void evaluate(VectorizedRowBatch batch) {
    if (childExpressions != null) {
        this.evaluateChildren(batch);
    }
    DecimalColumnVector inputColVector = (DecimalColumnVector) batch.cols[colNum];
    DecimalColumnVector outputColVector = (DecimalColumnVector) batch.cols[outputColumnNum];
    int[] sel = batch.selected;
    boolean[] inputIsNull = inputColVector.isNull;
    boolean[] outputIsNull = outputColVector.isNull;
    int n = batch.size;
    HiveDecimalWritable[] vector = inputColVector.vector;
    // return immediately if batch is empty
    if (n == 0) {
        return;
    }
    // We do not need to do a column reset since we are carefully changing the output.
    outputColVector.isRepeating = false;
    if (inputColVector.isRepeating) {
        if (inputColVector.noNulls || !inputIsNull[0]) {
            // Set isNull before call in case it changes it mind.
            outputIsNull[0] = false;
            round(0, vector[0], decimalPlaces, outputColVector);
        } else {
            outputIsNull[0] = true;
            outputColVector.noNulls = false;
        }
        outputColVector.isRepeating = true;
        return;
    }
    if (inputColVector.noNulls) {
        if (batch.selectedInUse) {
            if (!outputColVector.noNulls) {
                for (int j = 0; j != n; j++) {
                    final int i = sel[j];
                    // Set isNull before call in case it changes it mind.
                    outputIsNull[i] = false;
                    round(i, vector[i], decimalPlaces, outputColVector);
                }
            } else {
                for (int j = 0; j != n; j++) {
                    final int i = sel[j];
                    round(i, vector[i], decimalPlaces, outputColVector);
                }
            }
        } else {
            if (!outputColVector.noNulls) {
                // Assume it is almost always a performance win to fill all of isNull so we can
                // safely reset noNulls.
                Arrays.fill(outputIsNull, false);
                outputColVector.noNulls = true;
            }
            for (int i = 0; i != n; i++) {
                round(i, vector[i], decimalPlaces, outputColVector);
            }
        }
    } else /* there are nulls in the inputColVector */
    {
        // Carefully handle NULLs...
        outputColVector.noNulls = false;
        if (batch.selectedInUse) {
            for (int j = 0; j != n; j++) {
                int i = sel[j];
                outputIsNull[i] = inputIsNull[i];
                round(i, vector[i], decimalPlaces, outputColVector);
            }
        } else {
            System.arraycopy(inputIsNull, 0, outputIsNull, 0, n);
            for (int i = 0; i != n; i++) {
                round(i, vector[i], decimalPlaces, outputColVector);
            }
        }
    }
}

Example 82 with DecimalColumnVector

use of org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector in project hive by apache.

the class TestInputOutputFormat method testVectorizationWithAcid.

// test acid with vectorization, no combine
@Test
public void testVectorizationWithAcid() throws Exception {
    StructObjectInspector inspector = new BigRowInspector();
    JobConf conf = createMockExecutionEnvironment(workDir, new Path("mock:///"), "vectorizationAcid", inspector, true, 1);
    // write the orc file to the mock file system
    Path partDir = new Path(conf.get("mapred.input.dir"));
    OrcRecordUpdater writer = new OrcRecordUpdater(partDir, new AcidOutputFormat.Options(conf).maximumWriteId(10).writingBase(true).bucket(0).inspector(inspector).finalDestination(partDir));
    for (int i = 0; i < 100; ++i) {
        BigRow row = new BigRow(i);
        writer.insert(10, row);
    }
    writer.close(false);
    Path path = new Path("mock:/vectorizationAcid/p=0/base_0000010/bucket_00000");
    setBlocks(path, conf, new MockBlock("host0", "host1"));
    // call getsplits
    HiveInputFormat<?, ?> inputFormat = new HiveInputFormat<WritableComparable, Writable>();
    InputSplit[] splits = inputFormat.getSplits(conf, 10);
    assertEquals(1, splits.length);
    conf.set(IOConstants.SCHEMA_EVOLUTION_COLUMNS, BigRow.getColumnNamesProperty());
    conf.set(IOConstants.SCHEMA_EVOLUTION_COLUMNS_TYPES, BigRow.getColumnTypesProperty());
    HiveConf.setBoolVar(conf, HiveConf.ConfVars.HIVE_TRANSACTIONAL_TABLE_SCAN, true);
    org.apache.hadoop.mapred.RecordReader<NullWritable, VectorizedRowBatch> reader = inputFormat.getRecordReader(splits[0], conf, Reporter.NULL);
    NullWritable key = reader.createKey();
    VectorizedRowBatch value = reader.createValue();
    assertEquals(true, reader.next(key, value));
    assertEquals(100, value.count());
    LongColumnVector booleanColumn = (LongColumnVector) value.cols[0];
    LongColumnVector byteColumn = (LongColumnVector) value.cols[1];
    LongColumnVector shortColumn = (LongColumnVector) value.cols[2];
    LongColumnVector intColumn = (LongColumnVector) value.cols[3];
    LongColumnVector longColumn = (LongColumnVector) value.cols[4];
    DoubleColumnVector floatColumn = (DoubleColumnVector) value.cols[5];
    DoubleColumnVector doubleCoulmn = (DoubleColumnVector) value.cols[6];
    BytesColumnVector stringColumn = (BytesColumnVector) value.cols[7];
    DecimalColumnVector decimalColumn = (DecimalColumnVector) value.cols[8];
    LongColumnVector dateColumn = (LongColumnVector) value.cols[9];
    TimestampColumnVector timestampColumn = (TimestampColumnVector) value.cols[10];
    for (int i = 0; i < 100; i++) {
        assertEquals("checking boolean " + i, i % 2 == 0 ? 1 : 0, booleanColumn.vector[i]);
        assertEquals("checking byte " + i, (byte) i, byteColumn.vector[i]);
        assertEquals("checking short " + i, (short) i, shortColumn.vector[i]);
        assertEquals("checking int " + i, i, intColumn.vector[i]);
        assertEquals("checking long " + i, i, longColumn.vector[i]);
        assertEquals("checking float " + i, i, floatColumn.vector[i], 0.0001);
        assertEquals("checking double " + i, i, doubleCoulmn.vector[i], 0.0001);
        Text strValue = new Text();
        strValue.set(stringColumn.vector[i], stringColumn.start[i], stringColumn.length[i]);
        assertEquals("checking string " + i, new Text(Long.toHexString(i)), strValue);
        assertEquals("checking decimal " + i, HiveDecimal.create(i), decimalColumn.vector[i].getHiveDecimal());
        assertEquals("checking date " + i, i, dateColumn.vector[i]);
        long millis = (long) i * MILLIS_IN_DAY;
        millis -= LOCAL_TIMEZONE.getOffset(millis);
        assertEquals("checking timestamp " + i, millis, timestampColumn.getTime(i));
    }
    assertEquals(false, reader.next(key, value));
}

Example 83 with DecimalColumnVector

use of org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector in project hive by apache.

the class TestVectorTypeCasts method testCastDecimalToBoolean.

@Test
public /**
 * Just spot check the basic case because code path is the same as
 * for cast of decimal to long due to inheritance.
 */
void testCastDecimalToBoolean() throws HiveException {
    VectorizedRowBatch b = getBatchDecimalLong();
    VectorExpression expr = new CastDecimalToBoolean(0, 1);
    expr.setInputTypeInfos(new TypeInfo[] { TypeInfoFactory.decimalTypeInfo });
    expr.setOutputTypeInfo(TypeInfoFactory.booleanTypeInfo);
    expr.transientInit();
    DecimalColumnVector in = (DecimalColumnVector) b.cols[0];
    in.vector[1].set(HiveDecimal.create(0));
    expr.evaluate(b);
    LongColumnVector r = (LongColumnVector) b.cols[1];
    assertEquals(1, r.vector[0]);
    assertEquals(0, r.vector[1]);
    assertEquals(1, r.vector[2]);
}

Example 84 with DecimalColumnVector

use of org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector in project hive by apache.

the class TestVectorTypeCasts method getBatchLongDecimalPrec5Scale2.

/**
 * This batch has output decimal column precision 5 and scale 2.
 * The goal is to allow testing of input long values that, when
 * converted to decimal, will not fit in the given precision.
 * Then it will be possible to check that the results are NULL.
 */
private VectorizedRowBatch getBatchLongDecimalPrec5Scale2() {
    VectorizedRowBatch b = new VectorizedRowBatch(2);
    LongColumnVector lv;
    b.cols[0] = lv = new LongColumnVector();
    b.cols[1] = new DecimalColumnVector(5, 2);
    lv.vector[0] = 0;
    lv.vector[1] = -1;
    lv.vector[2] = 99999999999999L;
    return b;
}