Search in sources :

Example 46 with RowBatchReader

use of org.apache.drill.exec.physical.impl.scan.RowBatchReader in project drill by apache.

the class TestScanLifecycleLimit method testLimitOnSecondBatch.

/**
 * LIMIT 75, halfway through second batch.
 */
@Test
public void testLimitOnSecondBatch() {
    Pair<TwoReaderFactory, ScanLifecycle> pair = setupScan(75);
    TwoReaderFactory factory = pair.getLeft();
    ScanLifecycle scan = pair.getRight();
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    // First batch
    assertTrue(reader.next());
    RowSet result = fixture.wrap(reader.output());
    assertEquals(50, result.rowCount());
    result.clear();
    // Second batch
    assertTrue(reader.next());
    result = fixture.wrap(reader.output());
    assertEquals(25, result.rowCount());
    result.clear();
    // No third batch
    assertFalse(reader.next());
    reader.close();
    // No next reader, despite there being two, since we hit the limit.
    assertNull(scan.nextReader());
    scan.close();
    // Only the first of the two readers were created.
    assertEquals(1, factory.count());
}
Also used : RowBatchReader(org.apache.drill.exec.physical.impl.scan.RowBatchReader) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) Test(org.junit.Test)

Example 47 with RowBatchReader

use of org.apache.drill.exec.physical.impl.scan.RowBatchReader in project drill by apache.

the class TestScanLifecycleTwoReaders method testShrinkingSchemaWithConflict.

/**
 * Shrinking schema, as above. Explicit projection:<pre><code>
 * SELECT a, b FROM (a) then (a,b)
 * </code></pre><p>
 * But choose a missing column type (the default
 * Nullable INT) in the first reader that will conflict with the actual column type
 * (VARCHAR) in the second.
 */
@Test
public void testShrinkingSchemaWithConflict() {
    ScanLifecycleBuilder builder = new ScanLifecycleBuilder();
    builder.projection(RowSetTestUtils.projectList("a", "b"));
    builder.readerFactory(new TwoReaderFactory() {

        @Override
        public ManagedReader firstReader(SchemaNegotiator negotiator) {
            return new MockSingleColReader(negotiator);
        }

        @Override
        public ManagedReader secondReader(SchemaNegotiator negotiator) {
            return new MockEarlySchemaReader(negotiator, 1);
        }
    });
    ScanLifecycle scan = buildScan(builder);
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertTrue(reader.next());
    reader.output().clear();
    assertFalse(reader.next());
    reader.close();
    reader = scan.nextReader();
    try {
        reader.open();
        fail();
    } catch (UserException e) {
        assertTrue(e.getMessage().contains("conflict"));
    }
    reader.close();
    scan.close();
}
Also used : RowBatchReader(org.apache.drill.exec.physical.impl.scan.RowBatchReader) ManagedReader(org.apache.drill.exec.physical.impl.scan.v3.ManagedReader) SchemaNegotiator(org.apache.drill.exec.physical.impl.scan.v3.SchemaNegotiator) UserException(org.apache.drill.common.exceptions.UserException) ScanLifecycleBuilder(org.apache.drill.exec.physical.impl.scan.v3.ScanLifecycleBuilder) Test(org.junit.Test) EvfTest(org.apache.drill.categories.EvfTest)

Example 48 with RowBatchReader

use of org.apache.drill.exec.physical.impl.scan.RowBatchReader in project drill by apache.

the class TestScanLifecycleBasics method testEarlySchemaEmpty.

/**
 * Test SELECT * from an early-schema table of () (that is,
 * a schema that consists of zero columns.
 */
@Test
public void testEarlySchemaEmpty() {
    ScanLifecycleBuilder builder = new ScanLifecycleBuilder();
    builder.readerFactory(new SingleReaderFactory() {

        @Override
        public ManagedReader next(SchemaNegotiator negotiator) {
            return new MockEmptySchemaReader(negotiator);
        }
    });
    ScanLifecycle scan = buildScan(builder);
    assertSame(ProjectionType.ALL, scan.schemaTracker().projectionType());
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    // Early schema: so output schema is available after open
    TupleMetadata expectedSchema = new SchemaBuilder().build();
    assertEquals(expectedSchema, scan.outputSchema());
    assertTrue(reader.next());
    RowSet expected = fixture.rowSetBuilder(expectedSchema).addRow().addRow().build();
    RowSetUtilities.verify(expected, fixture.wrap(reader.output()));
    assertFalse(reader.next());
    reader.close();
    scan.close();
}
Also used : RowBatchReader(org.apache.drill.exec.physical.impl.scan.RowBatchReader) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) ManagedReader(org.apache.drill.exec.physical.impl.scan.v3.ManagedReader) SchemaBuilder(org.apache.drill.exec.record.metadata.SchemaBuilder) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) SchemaNegotiator(org.apache.drill.exec.physical.impl.scan.v3.SchemaNegotiator) ScanLifecycleBuilder(org.apache.drill.exec.physical.impl.scan.v3.ScanLifecycleBuilder) Test(org.junit.Test) EvfTest(org.apache.drill.categories.EvfTest)

Example 49 with RowBatchReader

use of org.apache.drill.exec.physical.impl.scan.RowBatchReader in project drill by apache.

the class TestScanLifecycleSchema method testDefinedSchemaSubset.

/**
 * The defined schema is a subset of the reader's schema; the
 * defined schema acts as a project list.
 */
@Test
public void testDefinedSchemaSubset() {
    ScanLifecycleBuilder builder = new ScanLifecycleBuilder();
    builder.definedSchema(SCHEMA);
    builder.readerFactory(new SingleReaderFactory() {

        @Override
        public ManagedReader next(SchemaNegotiator negotiator) {
            return new MockThreeColReader(negotiator);
        }
    });
    ScanLifecycle scan = buildScan(builder);
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertEquals(SCHEMA, scan.outputSchema());
    assertTrue(reader.next());
    RowSet expected = fixture.rowSetBuilder(SCHEMA).addRow(101, "wilma").addRow(102, "betty").build();
    RowSetUtilities.verify(expected, fixture.wrap(reader.output()));
    assertFalse(reader.next());
    reader.close();
    scan.close();
}
Also used : RowBatchReader(org.apache.drill.exec.physical.impl.scan.RowBatchReader) ManagedReader(org.apache.drill.exec.physical.impl.scan.v3.ManagedReader) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) SchemaNegotiator(org.apache.drill.exec.physical.impl.scan.v3.SchemaNegotiator) ScanLifecycleBuilder(org.apache.drill.exec.physical.impl.scan.v3.ScanLifecycleBuilder) Test(org.junit.Test) EvfTest(org.apache.drill.categories.EvfTest)

Example 50 with RowBatchReader

use of org.apache.drill.exec.physical.impl.scan.RowBatchReader in project drill by apache.

the class TestScanLifecycleSchema method testLenientProvidedSchemaSubset.

/**
 * Lenient provided schema which is a subset of the reader's schema; the
 * provided schema agrees with the reader types
 */
@Test
public void testLenientProvidedSchemaSubset() {
    ScanLifecycleBuilder builder = new ScanLifecycleBuilder();
    builder.providedSchema(SCHEMA);
    builder.readerFactory(new SingleReaderFactory() {

        @Override
        public ManagedReader next(SchemaNegotiator negotiator) {
            return new MockThreeColReader(negotiator);
        }
    });
    ScanLifecycle scan = buildScan(builder);
    assertSame(ProjectionType.ALL, scan.schemaTracker().projectionType());
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertEquals(MockThreeColReader.READER_SCHEMA, scan.outputSchema());
    assertTrue(reader.next());
    RowSet expected = fixture.rowSetBuilder(MockThreeColReader.READER_SCHEMA).addRow(101, "wilma", 1001).addRow(102, "betty", 1002).build();
    RowSetUtilities.verify(expected, fixture.wrap(reader.output()));
    assertFalse(reader.next());
    reader.close();
    scan.close();
}
Also used : RowBatchReader(org.apache.drill.exec.physical.impl.scan.RowBatchReader) ManagedReader(org.apache.drill.exec.physical.impl.scan.v3.ManagedReader) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) SchemaNegotiator(org.apache.drill.exec.physical.impl.scan.v3.SchemaNegotiator) ScanLifecycleBuilder(org.apache.drill.exec.physical.impl.scan.v3.ScanLifecycleBuilder) Test(org.junit.Test) EvfTest(org.apache.drill.categories.EvfTest)

Aggregations

RowBatchReader (org.apache.drill.exec.physical.impl.scan.RowBatchReader)51 Test (org.junit.Test)48 ManagedReader (org.apache.drill.exec.physical.impl.scan.v3.ManagedReader)43 EvfTest (org.apache.drill.categories.EvfTest)42 ScanLifecycleBuilder (org.apache.drill.exec.physical.impl.scan.v3.ScanLifecycleBuilder)37 SchemaNegotiator (org.apache.drill.exec.physical.impl.scan.v3.SchemaNegotiator)37 RowSet (org.apache.drill.exec.physical.rowSet.RowSet)28 SchemaBuilder (org.apache.drill.exec.record.metadata.SchemaBuilder)13 TupleMetadata (org.apache.drill.exec.record.metadata.TupleMetadata)13 UserException (org.apache.drill.common.exceptions.UserException)12 DummyFileWork (org.apache.drill.exec.physical.impl.scan.v3.file.BaseFileScanTest.DummyFileWork)6 BaseTestScanLifecycle (org.apache.drill.exec.physical.impl.scan.v3.lifecycle.BaseTestScanLifecycle)6 ScanLifecycle (org.apache.drill.exec.physical.impl.scan.v3.lifecycle.ScanLifecycle)6 SchemaPath (org.apache.drill.common.expression.SchemaPath)1 EarlyEofException (org.apache.drill.exec.physical.impl.scan.v3.ManagedReader.EarlyEofException)1 VectorContainer (org.apache.drill.exec.record.VectorContainer)1 Path (org.apache.hadoop.fs.Path)1