Search in sources :

Example 6 with RowBatchReader

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

the class TestScanLifecycleLimit method testLimitOnBatchEnd.

/**
 * LIMIT 50, same as batch size, to check boundary conditions.
 */
@Test
public void testLimitOnBatchEnd() {
    Pair<TwoReaderFactory, ScanLifecycle> pair = setupScan(50);
    TwoReaderFactory factory = pair.getLeft();
    ScanLifecycle scan = pair.getRight();
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertTrue(reader.next());
    RowSet result = fixture.wrap(reader.output());
    assertEquals(50, result.rowCount());
    result.clear();
    // No second 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 7 with RowBatchReader

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

the class TestScanLifecycleLimit method testLimit0.

/**
 * LIMIT 0, to obtain only the schema.
 */
@Test
public void testLimit0() {
    Pair<TwoReaderFactory, ScanLifecycle> pair = setupScan(0);
    TwoReaderFactory factory = pair.getLeft();
    ScanLifecycle scan = pair.getRight();
    // Reader builds schema, but returns no data, though the reader
    // itself is happy to provide data.
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertTrue(reader.next());
    RowSet result = fixture.wrap(reader.output());
    assertEquals(0, result.rowCount());
    assertEquals(1, result.schema().size());
    result.clear();
    // No second 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 8 with RowBatchReader

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

the class TestScanLifecycleLimit method testLimit1.

/**
 * LIMIT 1, simplest case
 */
@Test
public void testLimit1() {
    Pair<TwoReaderFactory, ScanLifecycle> pair = setupScan(1);
    TwoReaderFactory factory = pair.getLeft();
    ScanLifecycle scan = pair.getRight();
    // Reader builds schema, and stops after one row, though the reader
    // itself is happy to provide more.
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertTrue(reader.next());
    RowSet result = fixture.wrap(reader.output());
    assertEquals(1, result.rowCount());
    assertEquals(1, result.schema().size());
    result.clear();
    // No second 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 9 with RowBatchReader

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

the class TestScanLifecycleTwoReaders method testExpandingSchemaAllowingSchemaChange.

/**
 * SELECT * FROM two readers, one (a, b), the other (a, b, c).
 * With schema change enabled, the third column shows up only
 * in the second batch, forcing a schema change downstream.
 */
@Test
public void testExpandingSchemaAllowingSchemaChange() {
    ScanLifecycleBuilder builder = new ScanLifecycleBuilder();
    builder.allowSchemaChange(true);
    builder.readerFactory(new TwoReaderFactory() {

        @Override
        public ManagedReader firstReader(SchemaNegotiator negotiator) {
            return new MockLateSchemaReader(negotiator, 1);
        }

        @Override
        public ManagedReader secondReader(SchemaNegotiator negotiator) {
            return new MockThreeColReader(negotiator);
        }
    });
    ScanLifecycle scan = buildScan(builder);
    verifyStandardReader(scan, 0);
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    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)

Example 10 with RowBatchReader

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

the class TestScanLifecycleTwoReaders method doTestShrinkingSchema.

private void doTestShrinkingSchema(boolean allowSchemaChange) {
    ScanLifecycleBuilder builder = new ScanLifecycleBuilder();
    builder.allowSchemaChange(allowSchemaChange);
    builder.readerFactory(new TwoReaderFactory() {

        @Override
        public ManagedReader firstReader(SchemaNegotiator negotiator) {
            return new MockLateSchemaReader(negotiator, 1);
        }

        @Override
        public ManagedReader secondReader(SchemaNegotiator negotiator) {
            return new MockSingleColReader(negotiator);
        }
    });
    ScanLifecycle scan = buildScan(builder);
    verifyStandardReader(scan, 0);
    RowBatchReader reader = scan.nextReader();
    assertTrue(reader.open());
    assertTrue(reader.next());
    RowSet expected = fixture.rowSetBuilder(SCHEMA).addRow(101, null).addRow(102, null).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)

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