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

Example 51 with SingleRowSet

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

the class TestResultSetLoaderMaps method testMapAddition.

/**
 * Test adding a map to a loader after writing the first row.
 */
@Test
public void testMapAddition() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new OptionBuilder().setSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    assertEquals(1, rsLoader.schemaVersion());
    RowSetLoader rootWriter = rsLoader.writer();
    // Start without the map. Add a map after the first row.
    rsLoader.startBatch();
    rootWriter.addRow(10);
    int mapIndex = rootWriter.addColumn(SchemaBuilder.columnSchema("m", MinorType.MAP, DataMode.REQUIRED));
    TupleWriter mapWriter = rootWriter.tuple(mapIndex);
    // Add a column to the map with the same name as the top-level column.
    // Verifies that the name spaces are independent.
    mapWriter.addColumn(SchemaBuilder.columnSchema("a", MinorType.VARCHAR, DataMode.REQUIRED));
    rootWriter.addRow(20, objArray("fred")).addRow(30, objArray("barney"));
    RowSet actual = fixture.wrap(rsLoader.harvest());
    assertEquals(3, rsLoader.schemaVersion());
    assertEquals(3, actual.rowCount());
    // Validate first batch
    TupleMetadata expectedSchema = new SchemaBuilder().add("a", MinorType.INT).addMap("m").add("a", MinorType.VARCHAR).resumeSchema().buildSchema();
    SingleRowSet expected = fixture.rowSetBuilder(expectedSchema).addRow(10, objArray("")).addRow(20, objArray("fred")).addRow(30, objArray("barney")).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    rsLoader.close();
}

Example 52 with SingleRowSet

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

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() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addMap("m").addArray("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 some rows
    rsLoader.startBatch();
    rootWriter.addRow(10, objArray(intArray(110, 120, 130), strArray("d1.1", "d1.2", "d1.3", "d1.4"))).addRow(20, objArray(intArray(210), strArray())).addRow(30, objArray(intArray(), strArray("d3.1")));
    // Validate first batch
    RowSet actual = fixture.wrap(rsLoader.harvest());
    SingleRowSet expected = fixture.rowSetBuilder(schema).addRow(10, objArray(intArray(110, 120, 130), strArray("d1.1", "d1.2", "d1.3", "d1.4"))).addRow(20, objArray(intArray(210), strArray())).addRow(30, objArray(intArray(), strArray("d3.1"))).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    // Add another array after the first row in the second batch.
    rsLoader.startBatch();
    rootWriter.addRow(40, objArray(intArray(410, 420), strArray("d4.1", "d4.2"))).addRow(50, objArray(intArray(510), strArray("d5.1")));
    TupleWriter mapWriter = rootWriter.tuple("m");
    mapWriter.addColumn(SchemaBuilder.columnSchema("e", MinorType.VARCHAR, DataMode.REPEATED));
    rootWriter.addRow(60, objArray(intArray(610, 620), strArray("d6.1", "d6.2"), strArray("e6.1", "e6.2"))).addRow(70, objArray(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());
    // System.out.println(actual.schema().toString());
    expected = fixture.rowSetBuilder(actual.schema()).addRow(40, objArray(intArray(410, 420), strArray("d4.1", "d4.2"), strArray())).addRow(50, objArray(intArray(510), strArray("d5.1"), strArray())).addRow(60, objArray(intArray(610, 620), strArray("d6.1", "d6.2"), strArray("e6.1", "e6.2"))).addRow(70, objArray(intArray(710), strArray(), strArray("e7.1", "e7.2"))).build();
    // expected.print();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    rsLoader.close();
}

Example 53 with SingleRowSet

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

the class TestResultSetLoaderOmittedValues method testOmittedValuesAtEnd.

/**
 * Test "holes" in the middle of a batch, and unset columns at
 * the end. Ending the batch should fill in missing values.
 */
@Test
public void testOmittedValuesAtEnd() {
    // Create columns up front
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).add("b", MinorType.VARCHAR).addNullable("c", MinorType.VARCHAR).add("d", MinorType.INT).addNullable("e", MinorType.INT).addArray("f", MinorType.VARCHAR).buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new OptionBuilder().setSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    rsLoader.startBatch();
    int rowCount = 0;
    ScalarWriter arrayWriter;
    for (int i = 0; i < 2; i++) {
        // Row 0, 1
        rootWriter.start();
        rowCount++;
        rootWriter.scalar(0).setInt(rowCount);
        rootWriter.scalar(1).setString("b_" + rowCount);
        rootWriter.scalar(2).setString("c_" + rowCount);
        rootWriter.scalar(3).setInt(rowCount * 10);
        rootWriter.scalar(4).setInt(rowCount * 100);
        arrayWriter = rootWriter.column(5).array().scalar();
        arrayWriter.setString("f_" + rowCount + "-1");
        arrayWriter.setString("f_" + rowCount + "-2");
        rootWriter.save();
    }
    for (int i = 0; i < 2; i++) {
        // Rows 2, 3
        rootWriter.start();
        rowCount++;
        rootWriter.scalar(0).setInt(rowCount);
        rootWriter.scalar(1).setString("b_" + rowCount);
        rootWriter.scalar(3).setInt(rowCount * 10);
        arrayWriter = rootWriter.column(5).array().scalar();
        arrayWriter.setString("f_" + rowCount + "-1");
        arrayWriter.setString("f_" + rowCount + "-2");
        rootWriter.save();
    }
    for (int i = 0; i < 2; i++) {
        // Rows 4, 5
        rootWriter.start();
        rowCount++;
        rootWriter.scalar(0).setInt(rowCount);
        rootWriter.scalar(2).setString("c_" + rowCount);
        rootWriter.scalar(4).setInt(rowCount * 100);
        rootWriter.save();
    }
    for (int i = 0; i < 2; i++) {
        // Rows 6, 7
        rootWriter.start();
        rowCount++;
        rootWriter.scalar(0).setInt(rowCount);
        rootWriter.scalar(1).setString("b_" + rowCount);
        rootWriter.scalar(2).setString("c_" + rowCount);
        rootWriter.scalar(3).setInt(rowCount * 10);
        rootWriter.scalar(4).setInt(rowCount * 100);
        arrayWriter = rootWriter.column(5).array().scalar();
        arrayWriter.setString("f_" + rowCount + "-1");
        arrayWriter.setString("f_" + rowCount + "-2");
        rootWriter.save();
    }
    for (int i = 0; i < 2; i++) {
        // Rows 8, 9
        rootWriter.start();
        rowCount++;
        rootWriter.scalar(0).setInt(rowCount);
        rootWriter.save();
    }
    // Harvest the row and verify.
    RowSet actual = fixture.wrap(rsLoader.harvest());
    // actual.print();
    BatchSchema expectedSchema = new SchemaBuilder().add("a", MinorType.INT).add("b", MinorType.VARCHAR).addNullable("c", MinorType.VARCHAR).add("3", MinorType.INT).addNullable("e", MinorType.INT).addArray("f", MinorType.VARCHAR).build();
    SingleRowSet expected = fixture.rowSetBuilder(expectedSchema).addRow(1, "b_1", "c_1", 10, 100, strArray("f_1-1", "f_1-2")).addRow(2, "b_2", "c_2", 20, 200, strArray("f_2-1", "f_2-2")).addRow(3, "b_3", null, 30, null, strArray("f_3-1", "f_3-2")).addRow(4, "b_4", null, 40, null, strArray("f_4-1", "f_4-2")).addRow(5, "", "c_5", 0, 500, strArray()).addRow(6, "", "c_6", 0, 600, strArray()).addRow(7, "b_7", "c_7", 70, 700, strArray("f_7-1", "f_7-2")).addRow(8, "b_8", "c_8", 80, 800, strArray("f_8-1", "f_8-2")).addRow(9, "", null, 0, null, strArray()).addRow(10, "", null, 0, null, strArray()).build();
    new RowSetComparison(expected).verifyAndClearAll(actual);
    rsLoader.close();
}

Example 54 with SingleRowSet

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

the class TestResultSetLoaderOmittedValues method testSkipRows.

/**
 * Test that omitting the call to saveRow() effectively discards
 * the row. Note that the vectors still contain values in the
 * discarded position; just the various pointers are unset. If
 * the batch ends before the discarded values are overwritten, the
 * discarded values just exist at the end of the vector. Since vectors
 * start with garbage contents, the discarded values are simply a different
 * kind of garbage. But, if the client writes a new row, then the new
 * row overwrites the discarded row. This works because we only change
 * the tail part of a vector; never the internals.
 */
@Test
public void testSkipRows() {
    TupleMetadata schema = new SchemaBuilder().add("a", MinorType.INT).addNullable("b", MinorType.VARCHAR).buildSchema();
    ResultSetLoaderImpl.ResultSetOptions options = new OptionBuilder().setRowCountLimit(ValueVector.MAX_ROW_COUNT).setSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    rsLoader.startBatch();
    int rowNumber = 0;
    for (int i = 0; i < 14; i++) {
        rootWriter.start();
        rowNumber++;
        rootWriter.scalar(0).setInt(rowNumber);
        if (i % 3 == 0) {
            rootWriter.scalar(1).setNull();
        } else {
            rootWriter.scalar(1).setString("b-" + rowNumber);
        }
        if (i % 2 == 0) {
            rootWriter.save();
        }
    }
    RowSet result = fixture.wrap(rsLoader.harvest());
    // result.print();
    SingleRowSet expected = fixture.rowSetBuilder(result.batchSchema()).addRow(1, null).addRow(3, "b-3").addRow(5, "b-5").addRow(7, null).addRow(9, "b-9").addRow(11, "b-11").addRow(13, null).build();
    // expected.print();
    new RowSetComparison(expected).verifyAndClearAll(result);
    rsLoader.close();
}

Example 55 with SingleRowSet

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

the class TestResultSetLoaderProjection method testMapProjection.

@Test
public void testMapProjection() {
    List<SchemaPath> selection = Lists.newArrayList(SchemaPath.getSimplePath("m1"), SchemaPath.getCompoundPath("m2", "d"));
    TupleMetadata schema = new SchemaBuilder().addMap("m1").add("a", MinorType.INT).add("b", MinorType.INT).resumeSchema().addMap("m2").add("c", MinorType.INT).add("d", MinorType.INT).resumeSchema().addMap("m3").add("e", MinorType.INT).add("f", MinorType.INT).resumeSchema().buildSchema();
    ResultSetOptions options = new OptionBuilder().setProjection(selection).setSchema(schema).build();
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator(), options);
    RowSetLoader rootWriter = rsLoader.writer();
    // Verify the projected columns
    TupleMetadata actualSchema = rootWriter.schema();
    ColumnMetadata m1Md = actualSchema.metadata("m1");
    assertTrue(m1Md.isMap());
    assertTrue(m1Md.isProjected());
    assertEquals(2, m1Md.mapSchema().size());
    assertTrue(m1Md.mapSchema().metadata("a").isProjected());
    assertTrue(m1Md.mapSchema().metadata("b").isProjected());
    ColumnMetadata m2Md = actualSchema.metadata("m2");
    assertTrue(m2Md.isMap());
    assertTrue(m2Md.isProjected());
    assertEquals(2, m2Md.mapSchema().size());
    assertFalse(m2Md.mapSchema().metadata("c").isProjected());
    assertTrue(m2Md.mapSchema().metadata("d").isProjected());
    ColumnMetadata m3Md = actualSchema.metadata("m3");
    assertTrue(m3Md.isMap());
    assertFalse(m3Md.isProjected());
    assertEquals(2, m3Md.mapSchema().size());
    assertFalse(m3Md.mapSchema().metadata("e").isProjected());
    assertFalse(m3Md.mapSchema().metadata("f").isProjected());
    // Write a couple of rows.
    rsLoader.startBatch();
    rootWriter.start();
    rootWriter.tuple("m1").scalar("a").setInt(1);
    rootWriter.tuple("m1").scalar("b").setInt(2);
    rootWriter.tuple("m2").scalar("c").setInt(3);
    rootWriter.tuple("m2").scalar("d").setInt(4);
    rootWriter.tuple("m3").scalar("e").setInt(5);
    rootWriter.tuple("m3").scalar("f").setInt(6);
    rootWriter.save();
    rootWriter.start();
    rootWriter.tuple("m1").scalar("a").setInt(11);
    rootWriter.tuple("m1").scalar("b").setInt(12);
    rootWriter.tuple("m2").scalar("c").setInt(13);
    rootWriter.tuple("m2").scalar("d").setInt(14);
    rootWriter.tuple("m3").scalar("e").setInt(15);
    rootWriter.tuple("m3").scalar("f").setInt(16);
    rootWriter.save();
    // Verify. Only the projected columns appear in the result set.
    BatchSchema expectedSchema = new SchemaBuilder().addMap("m1").add("a", MinorType.INT).add("b", MinorType.INT).resumeSchema().addMap("m2").add("d", MinorType.INT).resumeSchema().build();
    SingleRowSet expected = fixture.rowSetBuilder(expectedSchema).addRow(objArray(1, 2), objArray(4)).addRow(objArray(11, 12), objArray(14)).build();
    new RowSetComparison(expected).verifyAndClearAll(fixture.wrap(rsLoader.harvest()));
    rsLoader.close();
}