Search in sources :

Example 11 with ScalarReader

use of org.apache.drill.exec.vector.accessor.ScalarReader in project drill by axbaretto.

the class RowSetTest method testRepeatedMapStructure.

@Test
public void testRepeatedMapStructure() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMapArray("m").add("b", MinorType.INT).add("c", MinorType.INT).resumeSchema().buildSchema();
    ExtendableRowSet rowSet = fixture.rowSet(schema);
    RowSetWriter writer = rowSet.writer();
    // Map and Int
    // Pick out components and lightly test. (Assumes structure
    // tested earlier is still valid, so no need to exhaustively
    // test again.)
    assertEquals(ObjectType.SCALAR, writer.column("a").type());
    assertEquals(ObjectType.ARRAY, writer.column("m").type());
    ArrayWriter maWriter = writer.column(1).array();
    assertEquals(ObjectType.TUPLE, maWriter.entryType());
    TupleWriter mapWriter = maWriter.tuple();
    assertEquals(ObjectType.SCALAR, mapWriter.column("b").type());
    assertEquals(ObjectType.SCALAR, mapWriter.column("c").type());
    ScalarWriter aWriter = writer.column("a").scalar();
    ScalarWriter bWriter = mapWriter.scalar("b");
    ScalarWriter cWriter = mapWriter.scalar("c");
    assertEquals(ValueType.INTEGER, aWriter.valueType());
    assertEquals(ValueType.INTEGER, bWriter.valueType());
    assertEquals(ValueType.INTEGER, cWriter.valueType());
    // Write data
    aWriter.setInt(10);
    bWriter.setInt(101);
    cWriter.setInt(102);
    // Advance to next array position
    maWriter.save();
    bWriter.setInt(111);
    cWriter.setInt(112);
    maWriter.save();
    writer.save();
    aWriter.setInt(20);
    bWriter.setInt(201);
    cWriter.setInt(202);
    maWriter.save();
    bWriter.setInt(211);
    cWriter.setInt(212);
    maWriter.save();
    writer.save();
    aWriter.setInt(30);
    bWriter.setInt(301);
    cWriter.setInt(302);
    maWriter.save();
    bWriter.setInt(311);
    cWriter.setInt(312);
    maWriter.save();
    writer.save();
    // Finish the row set and get a reader.
    SingleRowSet actual = writer.done();
    RowSetReader reader = actual.reader();
    // Verify reader structure
    assertEquals(ObjectType.SCALAR, reader.column("a").type());
    assertEquals(ObjectType.ARRAY, reader.column("m").type());
    ArrayReader maReader = reader.column(1).array();
    assertEquals(ObjectType.TUPLE, maReader.entryType());
    TupleReader mapReader = maReader.tuple();
    assertEquals(ObjectType.SCALAR, mapReader.column("b").type());
    assertEquals(ObjectType.SCALAR, mapReader.column("c").type());
    ScalarReader aReader = reader.column("a").scalar();
    ScalarReader bReader = mapReader.scalar("b");
    ScalarReader cReader = mapReader.scalar("c");
    assertEquals(ValueType.INTEGER, aReader.valueType());
    assertEquals(ValueType.INTEGER, bReader.valueType());
    assertEquals(ValueType.INTEGER, cReader.valueType());
    // Row 1: use index accessors
    assertTrue(reader.next());
    assertEquals(10, aReader.getInt());
    TupleReader ixReader = maReader.tuple(0);
    assertEquals(101, ixReader.scalar(0).getInt());
    assertEquals(102, ixReader.scalar(1).getInt());
    ixReader = maReader.tuple(1);
    assertEquals(111, ixReader.scalar(0).getInt());
    assertEquals(112, ixReader.scalar(1).getInt());
    // Row 2: use common accessor with explicit positioning,
    // but access scalars through the map reader.
    assertTrue(reader.next());
    assertEquals(20, aReader.getInt());
    maReader.setPosn(0);
    assertEquals(201, mapReader.scalar(0).getInt());
    assertEquals(202, mapReader.scalar(1).getInt());
    maReader.setPosn(1);
    assertEquals(211, mapReader.scalar(0).getInt());
    assertEquals(212, mapReader.scalar(1).getInt());
    // Row 3: use common accessor for scalars
    assertTrue(reader.next());
    assertEquals(30, aReader.getInt());
    maReader.setPosn(0);
    assertEquals(301, bReader.getInt());
    assertEquals(302, cReader.getInt());
    maReader.setPosn(1);
    assertEquals(311, bReader.getInt());
    assertEquals(312, cReader.getInt());
    assertFalse(reader.next());
    // Verify that the map accessor's value count was set.
    @SuppressWarnings("resource") RepeatedMapVector mapVector = (RepeatedMapVector) actual.container().getValueVector(1).getValueVector();
    assertEquals(3, mapVector.getAccessor().getValueCount());
    // Verify the readers and writers again using the testing tools.
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, objArray(objArray(101, 102), objArray(111, 112))).addRow(20, objArray(objArray(201, 202), objArray(211, 212))).addRow(30, objArray(objArray(301, 302), objArray(311, 312))).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
}
Also used : RowSetWriter(org.apache.drill.test.rowSet.RowSetWriter) TupleReader(org.apache.drill.exec.vector.accessor.TupleReader) SingleRowSet(org.apache.drill.test.rowSet.RowSet.SingleRowSet) RepeatedMapVector(org.apache.drill.exec.vector.complex.RepeatedMapVector) ScalarReader(org.apache.drill.exec.vector.accessor.ScalarReader) ArrayReader(org.apache.drill.exec.vector.accessor.ArrayReader) RowSetComparison(org.apache.drill.test.rowSet.RowSetComparison) TupleWriter(org.apache.drill.exec.vector.accessor.TupleWriter) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) SchemaBuilder(org.apache.drill.test.rowSet.schema.SchemaBuilder) ArrayWriter(org.apache.drill.exec.vector.accessor.ArrayWriter) RowSetReader(org.apache.drill.test.rowSet.RowSetReader) ScalarWriter(org.apache.drill.exec.vector.accessor.ScalarWriter) ExtendableRowSet(org.apache.drill.test.rowSet.RowSet.ExtendableRowSet) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 12 with ScalarReader

use of org.apache.drill.exec.vector.accessor.ScalarReader in project drill by axbaretto.

the class TestFillEmpties method doFillEmptiesScalar.

private void doFillEmptiesScalar(MajorType majorType) {
    TupleMetadata schema = new SchemaBuilder().add("a", majorType).buildSchema();
    ExtendableRowSet rs = fixture.rowSet(schema);
    RowSetWriter writer = rs.writer();
    ScalarWriter colWriter = writer.scalar(0);
    ValueType valueType = colWriter.valueType();
    boolean nullable = majorType.getMode() == DataMode.OPTIONAL;
    for (int i = 0; i < ROW_COUNT; i++) {
        if (i % 5 == 0) {
            colWriter.setObject(RowSetUtilities.testDataFromInt(valueType, majorType, i));
        }
        writer.save();
    }
    SingleRowSet result = writer.done();
    RowSetReader reader = result.reader();
    ScalarReader colReader = reader.scalar(0);
    MinorType type = majorType.getMinorType();
    boolean isVariable = (type == MinorType.VARCHAR || type == MinorType.VAR16CHAR || type == MinorType.VARBINARY);
    for (int i = 0; i < ROW_COUNT; i++) {
        assertTrue(reader.next());
        if (i % 5 != 0) {
            if (nullable) {
                // Nullable types fill with nulls.
                assertTrue(colReader.isNull());
                continue;
            }
            if (isVariable) {
                // Variable width types fill with a zero-length value.
                assertEquals(0, colReader.getBytes().length);
                continue;
            }
        }
        // All other types fill with zero-bytes, interpreted as some form
        // of zero for each type.
        Object actual = colReader.getObject();
        Object expected = RowSetUtilities.testDataFromInt(valueType, majorType, i % 5 == 0 ? i : 0);
        RowSetUtilities.assertEqualValues(majorType.toString().replace('\n', ' ') + "[" + i + "]", valueType, expected, actual);
    }
    result.clear();
}
Also used : RowSetWriter(org.apache.drill.test.rowSet.RowSetWriter) SingleRowSet(org.apache.drill.test.rowSet.RowSet.SingleRowSet) ValueType(org.apache.drill.exec.vector.accessor.ValueType) ScalarReader(org.apache.drill.exec.vector.accessor.ScalarReader) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) SchemaBuilder(org.apache.drill.test.rowSet.schema.SchemaBuilder) MinorType(org.apache.drill.common.types.TypeProtos.MinorType) RowSetReader(org.apache.drill.test.rowSet.RowSetReader) ScalarWriter(org.apache.drill.exec.vector.accessor.ScalarWriter) ExtendableRowSet(org.apache.drill.test.rowSet.RowSet.ExtendableRowSet)

Example 13 with ScalarReader

use of org.apache.drill.exec.vector.accessor.ScalarReader in project drill by axbaretto.

the class JsonFileBuilder method build.

public void build(File tableFile) throws IOException {
    tableFile.getParentFile().mkdirs();
    try (BufferedOutputStream os = new BufferedOutputStream(new FileOutputStream(tableFile))) {
        final RowSetReader reader = rowSet.reader();
        final int numCols = rowSet.batchSchema().getFieldCount();
        final Iterator<MaterializedField> fieldIterator = rowSet.batchSchema().iterator();
        final List<String> columnNames = Lists.newArrayList();
        final List<String> columnFormatters = Lists.newArrayList();
        // Build formatters from first row.
        while (fieldIterator.hasNext()) {
            final String columnName = fieldIterator.next().getName();
            final ScalarReader columnReader = reader.scalar(columnName);
            final ValueType valueType = columnReader.valueType();
            final String columnFormatter;
            if (customFormatters.containsKey(columnName)) {
                columnFormatter = customFormatters.get(columnName);
            } else if (DEFAULT_FORMATTERS.containsKey(valueType)) {
                columnFormatter = DEFAULT_FORMATTERS.get(valueType);
            } else {
                final String message = String.format("Unsupported column type %s", valueType);
                throw new UnsupportedOperationException(message);
            }
            columnNames.add(columnName);
            columnFormatters.add(columnFormatter);
        }
        final StringBuilder sb = new StringBuilder();
        String lineSeparator = "";
        for (int index = 0; index < rowSet.rowCount(); index++) {
            reader.next();
            sb.append(lineSeparator);
            sb.append('{');
            String separator = "";
            for (int columnIndex = 0; columnIndex < numCols; columnIndex++) {
                sb.append(separator);
                final String columnName = columnNames.get(columnIndex);
                final ScalarReader columnReader = reader.scalar(columnIndex);
                final String columnFormatter = columnFormatters.get(columnIndex);
                final Object columnObject = columnReader.getObject();
                final String columnString = String.format(columnFormatter, columnObject);
                sb.append('"').append(columnName).append('"').append(':').append(columnString);
                separator = ",";
            }
            sb.append('}');
            lineSeparator = "\n";
            os.write(sb.toString().getBytes());
            sb.delete(0, sb.length());
        }
    }
}
Also used : ValueType(org.apache.drill.exec.vector.accessor.ValueType) MaterializedField(org.apache.drill.exec.record.MaterializedField) ScalarReader(org.apache.drill.exec.vector.accessor.ScalarReader) FileOutputStream(java.io.FileOutputStream) RowSetReader(org.apache.drill.test.rowSet.RowSetReader) BufferedOutputStream(java.io.BufferedOutputStream)

Example 14 with ScalarReader

use of org.apache.drill.exec.vector.accessor.ScalarReader in project drill by apache.

the class TestResultSetLoaderRepeatedList method test2DOverflow.

@Test
public void test2DOverflow() {
    final TupleMetadata schema = new SchemaBuilder().add("id", MinorType.INT).addRepeatedList("list2").addArray(MinorType.VARCHAR).resumeSchema().buildSchema();
    final ResultSetLoaderImpl.ResultSetOptions options = new ResultSetOptionBuilder().rowCountLimit(ValueVector.MAX_ROW_COUNT).readerSchema(schema).build();
    final ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    final RowSetLoader writer = rsLoader.writer();
    // Fill the batch with enough data to cause overflow.
    // Data must be large enough to cause overflow before 64K rows
    // Make a bit bigger to overflow early.
    final int outerSize = 7;
    final int innerSize = 5;
    final int strLength = ValueVector.MAX_BUFFER_SIZE / ValueVector.MAX_ROW_COUNT / outerSize / innerSize + 20;
    final byte[] value = new byte[strLength - 6];
    Arrays.fill(value, (byte) 'X');
    final String strValue = new String(value, Charsets.UTF_8);
    int rowCount = 0;
    int elementCount = 0;
    final ArrayWriter outerWriter = writer.array(1);
    final ArrayWriter innerWriter = outerWriter.array();
    final ScalarWriter elementWriter = innerWriter.scalar();
    rsLoader.startBatch();
    while (!writer.isFull()) {
        writer.start();
        writer.scalar(0).setInt(rowCount);
        for (int j = 0; j < outerSize; j++) {
            for (int k = 0; k < innerSize; k++) {
                elementWriter.setString(String.format("%s%06d", strValue, elementCount));
                elementCount++;
            }
            outerWriter.save();
        }
        writer.save();
        rowCount++;
    }
    // Number of rows should be driven by vector size.
    // Our row count should include the overflow row
    final int expectedCount = ValueVector.MAX_BUFFER_SIZE / (strLength * innerSize * outerSize);
    assertEquals(expectedCount + 1, rowCount);
    // Loader's row count should include only "visible" rows
    assertEquals(expectedCount, writer.rowCount());
    // Total count should include invisible and look-ahead rows.
    assertEquals(expectedCount + 1, rsLoader.totalRowCount());
    // Result should exclude the overflow row
    RowSet result = fixture.wrap(rsLoader.harvest());
    assertEquals(expectedCount, result.rowCount());
    // Verify the data.
    RowSetReader reader = result.reader();
    ArrayReader outerReader = reader.array(1);
    ArrayReader innerReader = outerReader.array();
    ScalarReader strReader = innerReader.scalar();
    int readRowCount = 0;
    int readElementCount = 0;
    while (reader.next()) {
        assertEquals(readRowCount, reader.scalar(0).getInt());
        for (int i = 0; i < outerSize; i++) {
            assertTrue(outerReader.next());
            for (int j = 0; j < innerSize; j++) {
                assertTrue(innerReader.next());
                assertEquals(String.format("%s%06d", strValue, readElementCount), strReader.getString());
                readElementCount++;
            }
            assertFalse(innerReader.next());
        }
        assertFalse(outerReader.next());
        readRowCount++;
    }
    assertEquals(readRowCount, result.rowCount());
    result.clear();
    // Write a few more rows to verify the overflow row.
    rsLoader.startBatch();
    for (int i = 0; i < 1000; i++) {
        writer.start();
        writer.scalar(0).setInt(rowCount);
        for (int j = 0; j < outerSize; j++) {
            for (int k = 0; k < innerSize; k++) {
                elementWriter.setString(String.format("%s%06d", strValue, elementCount));
                elementCount++;
            }
            outerWriter.save();
        }
        writer.save();
        rowCount++;
    }
    result = fixture.wrap(rsLoader.harvest());
    assertEquals(1001, result.rowCount());
    final int startCount = readRowCount;
    reader = result.reader();
    outerReader = reader.array(1);
    innerReader = outerReader.array();
    strReader = innerReader.scalar();
    while (reader.next()) {
        assertEquals(readRowCount, reader.scalar(0).getInt());
        for (int i = 0; i < outerSize; i++) {
            assertTrue(outerReader.next());
            for (int j = 0; j < innerSize; j++) {
                assertTrue(innerReader.next());
                elementWriter.setString(String.format("%s%06d", strValue, readElementCount));
                assertEquals(String.format("%s%06d", strValue, readElementCount), strReader.getString());
                readElementCount++;
            }
            assertFalse(innerReader.next());
        }
        assertFalse(outerReader.next());
        readRowCount++;
    }
    assertEquals(readRowCount - startCount, result.rowCount());
    result.clear();
    rsLoader.close();
}
Also used : SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) ScalarReader(org.apache.drill.exec.vector.accessor.ScalarReader) ArrayReader(org.apache.drill.exec.vector.accessor.ArrayReader) ResultSetLoader(org.apache.drill.exec.physical.resultSet.ResultSetLoader) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) SchemaBuilder(org.apache.drill.exec.record.metadata.SchemaBuilder) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) ArrayWriter(org.apache.drill.exec.vector.accessor.ArrayWriter) RowSetReader(org.apache.drill.exec.physical.rowSet.RowSetReader) ScalarWriter(org.apache.drill.exec.vector.accessor.ScalarWriter) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 15 with ScalarReader

use of org.apache.drill.exec.vector.accessor.ScalarReader in project drill by apache.

the class TestResultSetLoaderMapArray method testDoubleNestedArray.

/**
 * Test a doubly-nested array of maps.
 */
@Test
public void testDoubleNestedArray() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMapArray("m1").add("b", MinorType.INT).addMapArray("m2").add("c", MinorType.INT).addArray("d", MinorType.VARCHAR).resumeMap().resumeSchema().buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new ResultSetOptionBuilder().readerSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    rsLoader.startBatch();
    ScalarWriter aWriter = rootWriter.scalar("a");
    ArrayWriter a1Writer = rootWriter.array("m1");
    TupleWriter m1Writer = a1Writer.tuple();
    ScalarWriter bWriter = m1Writer.scalar("b");
    ArrayWriter a2Writer = m1Writer.array("m2");
    TupleWriter m2Writer = a2Writer.tuple();
    ScalarWriter cWriter = m2Writer.scalar("c");
    ScalarWriter dWriter = m2Writer.array("d").scalar();
    for (int i = 0; i < 5; i++) {
        rootWriter.start();
        aWriter.setInt(i);
        for (int j = 0; j < 4; j++) {
            int a1Key = i + 10 + j;
            bWriter.setInt(a1Key);
            for (int k = 0; k < 3; k++) {
                int a2Key = a1Key * 10 + k;
                cWriter.setInt(a2Key);
                for (int l = 0; l < 2; l++) {
                    dWriter.setString("d-" + (a2Key * 10 + l));
                }
                a2Writer.save();
            }
            a1Writer.save();
        }
        rootWriter.save();
    }
    RowSet results = fixture.wrap(rsLoader.harvest());
    RowSetReader reader = results.reader();
    ScalarReader aReader = reader.scalar("a");
    ArrayReader a1Reader = reader.array("m1");
    TupleReader m1Reader = a1Reader.tuple();
    ScalarReader bReader = m1Reader.scalar("b");
    ArrayReader a2Reader = m1Reader.array("m2");
    TupleReader m2Reader = a2Reader.tuple();
    ScalarReader cReader = m2Reader.scalar("c");
    ArrayReader dArray = m2Reader.array("d");
    ScalarReader dReader = dArray.scalar();
    for (int i = 0; i < 5; i++) {
        assertTrue(reader.next());
        assertEquals(i, aReader.getInt());
        for (int j = 0; j < 4; j++) {
            assertTrue(a1Reader.next());
            int a1Key = i + 10 + j;
            assertEquals(a1Key, bReader.getInt());
            for (int k = 0; k < 3; k++) {
                assertTrue(a2Reader.next());
                int a2Key = a1Key * 10 + k;
                assertEquals(a2Key, cReader.getInt());
                for (int l = 0; l < 2; l++) {
                    assertTrue(dArray.next());
                    assertEquals("d-" + (a2Key * 10 + l), dReader.getString());
                }
            }
        }
    }
    rsLoader.close();
}
Also used : TupleReader(org.apache.drill.exec.vector.accessor.TupleReader) SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) ScalarReader(org.apache.drill.exec.vector.accessor.ScalarReader) ArrayReader(org.apache.drill.exec.vector.accessor.ArrayReader) ResultSetLoader(org.apache.drill.exec.physical.resultSet.ResultSetLoader) TupleWriter(org.apache.drill.exec.vector.accessor.TupleWriter) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) SchemaBuilder(org.apache.drill.exec.record.metadata.SchemaBuilder) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) ArrayWriter(org.apache.drill.exec.vector.accessor.ArrayWriter) RowSetReader(org.apache.drill.exec.physical.rowSet.RowSetReader) ScalarWriter(org.apache.drill.exec.vector.accessor.ScalarWriter) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Aggregations

ScalarReader (org.apache.drill.exec.vector.accessor.ScalarReader)97 SubOperatorTest (org.apache.drill.test.SubOperatorTest)78 Test (org.junit.Test)78 TupleMetadata (org.apache.drill.exec.record.metadata.TupleMetadata)72 SchemaBuilder (org.apache.drill.exec.record.metadata.SchemaBuilder)68 SingleRowSet (org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet)65 ScalarWriter (org.apache.drill.exec.vector.accessor.ScalarWriter)31 ArrayReader (org.apache.drill.exec.vector.accessor.ArrayReader)29 RowSetReader (org.apache.drill.test.rowSet.RowSetReader)27 SingleRowSet (org.apache.drill.test.rowSet.RowSet.SingleRowSet)26 SchemaBuilder (org.apache.drill.test.rowSet.schema.SchemaBuilder)26 BatchSchema (org.apache.drill.exec.record.BatchSchema)22 ExtendableRowSet (org.apache.drill.exec.physical.rowSet.RowSet.ExtendableRowSet)19 Period (org.joda.time.Period)15 ArrayWriter (org.apache.drill.exec.vector.accessor.ArrayWriter)13 TupleReader (org.apache.drill.exec.vector.accessor.TupleReader)10 TupleWriter (org.apache.drill.exec.vector.accessor.TupleWriter)10 VectorContainer (org.apache.drill.exec.record.VectorContainer)9 BigDecimal (java.math.BigDecimal)7 MajorType (org.apache.drill.common.types.TypeProtos.MajorType)7