Examples with RowSet org.apache.drill.test.rowSet.RowSet used on opensource projects

Example 6 with RowSet

use of org.apache.drill.test.rowSet.RowSet in project drill by axbaretto.

the class AbstractGenericCopierTest method testCopyRecords.

@Test
public void testCopyRecords() throws SchemaChangeException {
    try (RootAllocator allocator = new RootAllocator(10_000_000)) {
        final BatchSchema batchSchema = createTestSchema(BatchSchema.SelectionVectorMode.NONE);
        final RowSet srcRowSet = createSrcRowSet(allocator);
        final RowSet destRowSet = new RowSetBuilder(allocator, batchSchema).build();
        final VectorContainer destContainer = destRowSet.container();
        final Copier copier = createCopier();
        final RowSet expectedRowSet = createExpectedRowset(allocator);
        copier.setup(new RowSetBatch(srcRowSet), destContainer);
        copier.copyRecords(0, 3);
        try {
            new RowSetComparison(expectedRowSet).verify(destRowSet);
        } finally {
            srcRowSet.clear();
            if (srcRowSet instanceof RowSet.HyperRowSet) {
                ((RowSet.HyperRowSet) srcRowSet).getSv4().clear();
            }
            destRowSet.clear();
            expectedRowSet.clear();
        }
    }
}

Example 7 with RowSet

use of org.apache.drill.test.rowSet.RowSet in project drill by axbaretto.

the class TestSortImpl method runLargeSortTest.

/**
 * Run a full-blown sort test with multiple input batches. Because we want to
 * generate multiple inputs, we don't create them statically. Instead, we generate
 * them on the fly using a data generator. A matching data validator verifies the
 * output. Here, we are focusing on overall test flow. Separate, detailed, unit
 * tests have already probed the details of each sort component and data type,
 * so we don't need to repeat that whole exercise here; using integer keys is
 * sufficient.
 *
 * @param fixture the operator test fixture
 * @param dataGen input batch generator
 * @param validator validates output batches
 */
public void runLargeSortTest(OperatorFixture fixture, DataGenerator dataGen, DataValidator validator) {
    VectorContainer dest = new VectorContainer();
    SortImpl sort = makeSortImpl(fixture, Ordering.ORDER_ASC, Ordering.NULLS_UNSPECIFIED, dest);
    int batchCount = 0;
    RowSet input;
    while ((input = dataGen.nextRowSet()) != null) {
        batchCount++;
        if (batchCount == 1) {
            // Simulates a NEW_SCHEMA event
            sort.setSchema(input.container().getSchema());
        }
        // Simulates an OK event
        sort.addBatch(input.vectorAccessible());
    }
    // Simulate returning results
    SortResults results = sort.startMerge();
    if (results.getContainer() != dest) {
        dest.clear();
        dest = results.getContainer();
    }
    while (results.next()) {
        RowSet output = toRowSet(results, dest);
        validator.validate(output);
    }
    validator.validateDone();
    results.close();
    dest.clear();
    sort.close();
    sort.opContext().close();
}

Example 8 with RowSet

use of org.apache.drill.test.rowSet.RowSet in project drill by axbaretto.

the class TestResultSetLoaderMapArray method testBasics.

@Test
public void testBasics() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMapArray("m").add("c", MinorType.INT).add("d", MinorType.VARCHAR).resumeSchema().buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new OptionBuilder().setSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    // Verify structure and schema
    TupleMetadata actualSchema = rootWriter.schema();
    assertEquals(2, actualSchema.size());
    assertTrue(actualSchema.metadata(1).isArray());
    assertTrue(actualSchema.metadata(1).isMap());
    assertEquals(2, actualSchema.metadata("m").mapSchema().size());
    assertEquals(2, actualSchema.column("m").getChildren().size());
    // Write a couple of rows with arrays.
    rsLoader.startBatch();
    rootWriter.addRow(10, objArray(objArray(110, "d1.1"), objArray(120, "d2.2"))).addRow(20, objArray()).addRow(30, objArray(objArray(310, "d3.1"), objArray(320, "d3.2"), objArray(330, "d3.3")));
    // Verify the first batch
    RowSet actual = fixture.wrap(rsLoader.harvest());
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, objArray(objArray(110, "d1.1"), objArray(120, "d2.2"))).addRow(20, objArray()).addRow(30, objArray(objArray(310, "d3.1"), objArray(320, "d3.2"), objArray(330, "d3.3"))).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    // In the second, create a row, then add a map member.
    // Should be back-filled to empty for the first row.
    rsLoader.startBatch();
    rootWriter.addRow(40, objArray(objArray(410, "d4.1"), objArray(420, "d4.2")));
    TupleWriter mapWriter = rootWriter.array("m").tuple();
    mapWriter.addColumn(SchemaBuilder.columnSchema("e", MinorType.VARCHAR, DataMode.OPTIONAL));
    rootWriter.addRow(50, objArray(objArray(510, "d5.1", "e5.1"), objArray(520, "d5.2", null))).addRow(60, objArray(objArray(610, "d6.1", "e6.1"), objArray(620, "d6.2", null), objArray(630, "d6.3", "e6.3")));
    // Verify the second batch
    actual = fixture.wrap(rsLoader.harvest());
    TupleMetadata expectedSchema = new SchemaBuilder().add("a", MinorType.INT).addMapArray("m").add("c", MinorType.INT).add("d", MinorType.VARCHAR).addNullable("e", MinorType.VARCHAR).resumeSchema().buildSchema();
    expected = fixture.rowSetBuilder(expectedSchema).addRow(40, objArray(objArray(410, "d4.1", null), objArray(420, "d4.2", null))).addRow(50, objArray(objArray(510, "d5.1", "e5.1"), objArray(520, "d5.2", null))).addRow(60, objArray(objArray(610, "d6.1", "e6.1"), objArray(620, "d6.2", null), objArray(630, "d6.3", "e6.3"))).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    rsLoader.close();
}

Example 9 with RowSet

use of org.apache.drill.test.rowSet.RowSet in project drill by axbaretto.

the class TestResultSetLoaderMapArray method testNestedArray.

@Test
public void testNestedArray() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMapArray("m").add("c", MinorType.INT).addArray("d", MinorType.VARCHAR).resumeSchema().buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new OptionBuilder().setSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    // Write a couple of rows with arrays within arrays.
    // (And, of course, the Varchar is actually an array of
    // bytes, so that's three array levels.)
    rsLoader.startBatch();
    rootWriter.addRow(10, objArray(objArray(110, strArray("d1.1.1", "d1.1.2")), objArray(120, strArray("d1.2.1", "d1.2.2")))).addRow(20, objArray()).addRow(30, objArray(objArray(310, strArray("d3.1.1", "d3.2.2")), objArray(320, strArray()), objArray(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(objArray(110, strArray("d1.1.1", "d1.1.2")), objArray(120, strArray("d1.2.1", "d1.2.2")))).addRow(20, objArray()).addRow(30, objArray(objArray(310, strArray("d3.1.1", "d3.2.2")), objArray(320, strArray()), objArray(330, strArray("d3.3.1", "d1.2.2")))).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    rsLoader.close();
}

Example 10 with RowSet

use of org.apache.drill.test.rowSet.RowSet in project drill by axbaretto.

the class TestResultSetLoaderMapArray method testOverwriteRow.

/**
 * Version of the {#link TestResultSetLoaderProtocol#testOverwriteRow()} test
 * that uses nested columns inside an array of maps. Here we must call
 * <tt>start()</tt> to reset the array back to the initial start position after
 * each "discard."
 */
@Test
public void testOverwriteRow() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMapArray("m").add("b", MinorType.INT).add("c", MinorType.VARCHAR).resumeSchema().buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new OptionBuilder().setSchema(schema).setRowCountLimit(ValueVector.MAX_ROW_COUNT).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    // Can't use the shortcut to populate rows when doing overwrites.
    ScalarWriter aWriter = rootWriter.scalar("a");
    ArrayWriter maWriter = rootWriter.array("m");
    TupleWriter mWriter = maWriter.tuple();
    ScalarWriter bWriter = mWriter.scalar("b");
    ScalarWriter cWriter = mWriter.scalar("c");
    // Write 100,000 rows, overwriting 99% of them. This will cause vector
    // overflow and data corruption if overwrite does not work; but will happily
    // produce the correct result if everything works as it should.
    byte[] value = new byte[512];
    Arrays.fill(value, (byte) 'X');
    int count = 0;
    rsLoader.startBatch();
    while (count < 10_000) {
        rootWriter.start();
        count++;
        aWriter.setInt(count);
        for (int i = 0; i < 10; i++) {
            bWriter.setInt(count * 10 + i);
            cWriter.setBytes(value, value.length);
            maWriter.save();
        }
        if (count % 100 == 0) {
            rootWriter.save();
        }
    }
    // Verify using a reader.
    RowSet result = fixture.wrap(rsLoader.harvest());
    assertEquals(count / 100, result.rowCount());
    RowSetReader reader = result.reader();
    ArrayReader maReader = reader.array("m");
    TupleReader mReader = maReader.tuple();
    int rowId = 1;
    while (reader.next()) {
        assertEquals(rowId * 100, reader.scalar("a").getInt());
        assertEquals(10, maReader.size());
        for (int i = 0; i < 10; i++) {
            maReader.setPosn(i);
            assertEquals(rowId * 1000 + i, mReader.scalar("b").getInt());
            assertTrue(Arrays.equals(value, mReader.scalar("c").getBytes()));
        }
        rowId++;
    }
    result.clear();
    rsLoader.close();
}