Search in sources :

Example 61 with RowSetLoader

use of org.apache.drill.exec.physical.resultSet.RowSetLoader in project drill by apache.

the class TestResultSetLoaderDictArray method testArrayValue.

@Test
public void testArrayValue() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addDictArray("d", MinorType.INT).repeatedValue(MinorType.VARCHAR).resumeSchema().buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new ResultSetOptionBuilder().readerSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    // Write a couple of rows
    rsLoader.startBatch();
    rootWriter.addRow(10, objArray(map(110, strArray("d1.1.1", "d1.1.2"), 111, strArray("d1.1.3", "d1.1.4"), 112, strArray("d1.1.5", "d1.1.6")), map(120, strArray("d1.2.1", "d1.2.2")))).addRow(20, objArray()).addRow(30, objArray(map(310, strArray("d3.1.1", "d3.2.2"), 311, strArray("d3.1.3", "d3.2.4", "d3.1.5", "d3.1.6")), map(320, strArray(), 321, strArray("d3.2.2")), map(330, strArray("d3.3.1", "d1.2.2"))));
    // Verify the batch
    RowSet actual = fixture.wrap(rsLoader.harvest());
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, objArray(map(110, strArray("d1.1.1", "d1.1.2"), 111, strArray("d1.1.3", "d1.1.4"), 112, strArray("d1.1.5", "d1.1.6")), map(120, strArray("d1.2.1", "d1.2.2")))).addRow(20, objArray()).addRow(30, objArray(map(310, strArray("d3.1.1", "d3.2.2"), 311, strArray("d3.1.3", "d3.2.4", "d3.1.5", "d3.1.6")), map(320, strArray(), 321, strArray("d3.2.2")), map(330, strArray("d3.3.1", "d1.2.2")))).build();
    RowSetUtilities.verify(expected, actual);
    rsLoader.close();
}
Also used : SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) ResultSetLoader(org.apache.drill.exec.physical.resultSet.ResultSetLoader) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) SchemaBuilder(org.apache.drill.exec.record.metadata.SchemaBuilder) SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 62 with RowSetLoader

use of org.apache.drill.exec.physical.resultSet.RowSetLoader in project drill by apache.

the class TestResultSetLoaderMaps method testNestedMapsNullable.

/**
 * Create nested maps. Then, add columns to each map
 * on the fly. This time, with nullable types.
 */
@Test
public void testNestedMapsNullable() {
    final TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMap("m1").addNullable("b", MinorType.VARCHAR).addMap("m2").addNullable("c", MinorType.VARCHAR).resumeMap().resumeSchema().buildSchema();
    final ResultSetLoaderImpl.ResultSetOptions options = new ResultSetOptionBuilder().readerSchema(schema).build();
    final ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    final RowSetLoader rootWriter = rsLoader.writer();
    rsLoader.startBatch();
    rootWriter.addRow(10, mapValue("b1", mapValue("c1")));
    // Validate first batch
    RowSet actual = fixture.wrap(rsLoader.harvest());
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, mapValue("b1", mapValue("c1"))).build();
    RowSetUtilities.verify(expected, actual);
    // Now add columns in the second batch.
    rsLoader.startBatch();
    rootWriter.addRow(20, mapValue("b2", mapValue("c2")));
    final TupleWriter m1Writer = rootWriter.tuple("m1");
    m1Writer.addColumn(SchemaBuilder.columnSchema("d", MinorType.VARCHAR, DataMode.OPTIONAL));
    final TupleWriter m2Writer = m1Writer.tuple("m2");
    m2Writer.addColumn(SchemaBuilder.columnSchema("e", MinorType.VARCHAR, DataMode.OPTIONAL));
    rootWriter.addRow(30, mapValue("b3", mapValue("c3", "e3"), "d3"));
    // And another set while the write proceeds.
    m1Writer.addColumn(SchemaBuilder.columnSchema("f", MinorType.VARCHAR, DataMode.OPTIONAL));
    m2Writer.addColumn(SchemaBuilder.columnSchema("g", MinorType.VARCHAR, DataMode.OPTIONAL));
    rootWriter.addRow(40, mapValue("b4", mapValue("c4", "e4", "g4"), "d4", "e4"));
    // Validate second batch
    actual = fixture.wrap(rsLoader.harvest());
    final TupleMetadata expectedSchema = new SchemaBuilder().add("a", MinorType.INT).addMap("m1").addNullable("b", MinorType.VARCHAR).addMap("m2").addNullable("c", MinorType.VARCHAR).addNullable("e", MinorType.VARCHAR).addNullable("g", MinorType.VARCHAR).resumeMap().addNullable("d", MinorType.VARCHAR).addNullable("f", MinorType.VARCHAR).resumeSchema().buildSchema();
    expected = fixture.rowSetBuilder(expectedSchema).addRow(20, mapValue("b2", mapValue("c2", null, null), null, null)).addRow(30, mapValue("b3", mapValue("c3", "e3", null), "d3", null)).addRow(40, mapValue("b4", mapValue("c4", "e4", "g4"), "d4", "e4")).build();
    RowSetUtilities.verify(expected, actual);
    rsLoader.close();
}
Also used : SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) 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) SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 63 with RowSetLoader

use of org.apache.drill.exec.physical.resultSet.RowSetLoader in project drill by apache.

the class TestResultSetLoaderMaps method testMapWithArray.

/**
 * Test a map that contains a scalar array. No reason to suspect that this
 * will have problem as the array writer is fully tested in the accessor
 * subsystem. Still, need to test the cardinality methods of the loader
 * layer.
 */
@Test
public void testMapWithArray() {
    final TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMap("m").addArray("c", MinorType.INT).addArray("d", MinorType.VARCHAR).resumeSchema().buildSchema();
    final ResultSetLoaderImpl.ResultSetOptions options = new ResultSetOptionBuilder().readerSchema(schema).build();
    final ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    final RowSetLoader rootWriter = rsLoader.writer();
    // Write some rows
    rsLoader.startBatch();
    rootWriter.addRow(10, mapValue(intArray(110, 120, 130), strArray("d1.1", "d1.2", "d1.3", "d1.4"))).addRow(20, mapValue(intArray(210), strArray())).addRow(30, mapValue(intArray(), strArray("d3.1")));
    // Validate first batch
    RowSet actual = fixture.wrap(rsLoader.harvest());
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, mapValue(intArray(110, 120, 130), strArray("d1.1", "d1.2", "d1.3", "d1.4"))).addRow(20, mapValue(intArray(210), strArray())).addRow(30, mapValue(intArray(), strArray("d3.1"))).build();
    RowSetUtilities.verify(expected, actual);
    // Add another array after the first row in the second batch.
    rsLoader.startBatch();
    rootWriter.addRow(40, mapValue(intArray(410, 420), strArray("d4.1", "d4.2"))).addRow(50, mapValue(intArray(510), strArray("d5.1")));
    final TupleWriter mapWriter = rootWriter.tuple("m");
    mapWriter.addColumn(SchemaBuilder.columnSchema("e", MinorType.VARCHAR, DataMode.REPEATED));
    rootWriter.addRow(60, mapValue(intArray(610, 620), strArray("d6.1", "d6.2"), strArray("e6.1", "e6.2"))).addRow(70, mapValue(intArray(710), strArray(), strArray("e7.1", "e7.2")));
    // Validate first batch. The new array should have been back-filled with
    // empty offsets for the missing rows.
    actual = fixture.wrap(rsLoader.harvest());
    expected = fixture.rowSetBuilder(actual.schema()).addRow(40, mapValue(intArray(410, 420), strArray("d4.1", "d4.2"), strArray())).addRow(50, mapValue(intArray(510), strArray("d5.1"), strArray())).addRow(60, mapValue(intArray(610, 620), strArray("d6.1", "d6.2"), strArray("e6.1", "e6.2"))).addRow(70, mapValue(intArray(710), strArray(), strArray("e7.1", "e7.2"))).build();
    RowSetUtilities.verify(expected, actual);
    rsLoader.close();
}
Also used : SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) 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) SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 64 with RowSetLoader

use of org.apache.drill.exec.physical.resultSet.RowSetLoader in project drill by apache.

the class TestResultSetLoaderMaps method testMapEvolution.

/**
 * Create schema with a map, then add columns to the map
 * after delivering the first batch. The new columns should appear
 * in the second-batch output.
 */
@Test
public void testMapEvolution() {
    final TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMap("m").add("b", MinorType.VARCHAR).resumeSchema().buildSchema();
    final ResultSetLoaderImpl.ResultSetOptions options = new ResultSetOptionBuilder().readerSchema(schema).build();
    final ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    assertEquals(3, rsLoader.schemaVersion());
    final RowSetLoader rootWriter = rsLoader.writer();
    rsLoader.startBatch();
    rootWriter.addRow(10, mapValue("fred")).addRow(20, mapValue("barney"));
    RowSet actual = fixture.wrap(rsLoader.harvest());
    assertEquals(3, rsLoader.schemaVersion());
    assertEquals(2, actual.rowCount());
    // Validate first batch
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, mapValue("fred")).addRow(20, mapValue("barney")).build();
    RowSetUtilities.verify(expected, actual);
    // Add three columns in the second batch. One before
    // the batch starts, one before the first row, and one after
    // the first row.
    final TupleWriter mapWriter = rootWriter.tuple("m");
    mapWriter.addColumn(SchemaBuilder.columnSchema("c", MinorType.INT, DataMode.REQUIRED));
    rsLoader.startBatch();
    mapWriter.addColumn(SchemaBuilder.columnSchema("d", MinorType.BIGINT, DataMode.REQUIRED));
    rootWriter.addRow(30, mapValue("wilma", 130, 130_000L));
    mapWriter.addColumn(SchemaBuilder.columnSchema("e", MinorType.VARCHAR, DataMode.REQUIRED));
    rootWriter.addRow(40, mapValue("betty", 140, 140_000L, "bam-bam"));
    actual = fixture.wrap(rsLoader.harvest());
    assertEquals(6, rsLoader.schemaVersion());
    assertEquals(2, actual.rowCount());
    // Validate first batch
    final TupleMetadata expectedSchema = new SchemaBuilder().add("a", MinorType.INT).addMap("m").add("b", MinorType.VARCHAR).add("c", MinorType.INT).add("d", MinorType.BIGINT).add("e", MinorType.VARCHAR).resumeSchema().buildSchema();
    expected = fixture.rowSetBuilder(expectedSchema).addRow(30, mapValue("wilma", 130, 130_000L, "")).addRow(40, mapValue("betty", 140, 140_000L, "bam-bam")).build();
    RowSetUtilities.verify(expected, actual);
    rsLoader.close();
}
Also used : SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) 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) SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 65 with RowSetLoader

use of org.apache.drill.exec.physical.resultSet.RowSetLoader in project drill by apache.

the class TestResultSetLoaderOverflow method testVectorSizeLimitWithAppend.

@Test
public void testVectorSizeLimitWithAppend() {
    TupleMetadata schema = new SchemaBuilder().add("s", MinorType.VARCHAR).buildSchema();
    ResultSetOptions options = new ResultSetOptionBuilder().rowCountLimit(ValueVector.MAX_ROW_COUNT).readerSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    rsLoader.startBatch();
    byte[] head = "abc".getBytes();
    byte[] tail = new byte[523];
    Arrays.fill(tail, (byte) 'X');
    String expected = new String(head, Charsets.UTF_8);
    expected += new String(tail, Charsets.UTF_8);
    expected += new String(tail, Charsets.UTF_8);
    int count = 0;
    ScalarWriter colWriter = rootWriter.scalar(0);
    while (!rootWriter.isFull()) {
        rootWriter.start();
        colWriter.setBytes(head, head.length);
        colWriter.appendBytes(tail, tail.length);
        colWriter.appendBytes(tail, tail.length);
        rootWriter.save();
        count++;
    }
    // Number of rows should be driven by vector size.
    // Our row count should include the overflow row
    int valueLength = head.length + 2 * tail.length;
    int expectedCount = ValueVector.MAX_BUFFER_SIZE / valueLength;
    assertEquals(expectedCount + 1, count);
    // Loader's row count should include only "visible" rows
    assertEquals(expectedCount, rootWriter.rowCount());
    // Total count should include invisible and look-ahead rows.
    assertEquals(expectedCount + 1, rsLoader.totalRowCount());
    // Result should exclude the overflow row
    {
        VectorContainer container = rsLoader.harvest();
        BatchValidator.validate(container);
        RowSet result = fixture.wrap(container);
        assertEquals(expectedCount, result.rowCount());
        // Verify that the values were, in fact, appended.
        RowSetReader reader = result.reader();
        while (reader.next()) {
            assertEquals(expected, reader.scalar(0).getString());
        }
        result.clear();
    }
    // Next batch should start with the overflow row
    rsLoader.startBatch();
    assertEquals(1, rootWriter.rowCount());
    assertEquals(expectedCount + 1, rsLoader.totalRowCount());
    {
        VectorContainer container = rsLoader.harvest();
        BatchValidator.validate(container);
        RowSet result = fixture.wrap(container);
        assertEquals(1, result.rowCount());
        RowSetReader reader = result.reader();
        while (reader.next()) {
            assertEquals(expected, reader.scalar(0).getString());
        }
        result.clear();
    }
    rsLoader.close();
}
Also used : RowSet(org.apache.drill.exec.physical.rowSet.RowSet) ResultSetOptions(org.apache.drill.exec.physical.resultSet.impl.ResultSetLoaderImpl.ResultSetOptions) VectorContainer(org.apache.drill.exec.record.VectorContainer) 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) 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

RowSetLoader (org.apache.drill.exec.physical.resultSet.RowSetLoader)98 ResultSetLoader (org.apache.drill.exec.physical.resultSet.ResultSetLoader)90 Test (org.junit.Test)86 SubOperatorTest (org.apache.drill.test.SubOperatorTest)85 SchemaBuilder (org.apache.drill.exec.record.metadata.SchemaBuilder)82 TupleMetadata (org.apache.drill.exec.record.metadata.TupleMetadata)82 SingleRowSet (org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet)66 RowSet (org.apache.drill.exec.physical.rowSet.RowSet)63 ScalarWriter (org.apache.drill.exec.vector.accessor.ScalarWriter)25 TupleWriter (org.apache.drill.exec.vector.accessor.TupleWriter)25 ResultSetOptions (org.apache.drill.exec.physical.resultSet.impl.ResultSetLoaderImpl.ResultSetOptions)23 RowSetReader (org.apache.drill.exec.physical.rowSet.RowSetReader)17 ArrayWriter (org.apache.drill.exec.vector.accessor.ArrayWriter)16 VectorContainer (org.apache.drill.exec.record.VectorContainer)15 SchemaPath (org.apache.drill.common.expression.SchemaPath)12 DictWriter (org.apache.drill.exec.vector.accessor.DictWriter)11 EvfTest (org.apache.drill.categories.EvfTest)10 MaterializedField (org.apache.drill.exec.record.MaterializedField)9 ColumnMetadata (org.apache.drill.exec.record.metadata.ColumnMetadata)6 ArrayReader (org.apache.drill.exec.vector.accessor.ArrayReader)5