Search in sources :

Example 76 with ColumnMetadata

use of org.apache.drill.exec.record.metadata.ColumnMetadata in project drill by apache.

the class TestResultSetLoaderProtocol method testCaseInsensitiveSchema.

/**
 * Schemas are case insensitive by default. Verify that
 * the schema mechanism works, with emphasis on the
 * case insensitive case.
 * <p>
 * The tests here and elsewhere build columns from a
 * <tt>MaterializedField</tt>. Doing so is rather old-school;
 * better to use the newer <tt>ColumnMetadata</tt> which provides
 * additional information. The code here simply uses the <tt>MaterializedField</tt>
 * to create a <tt>ColumnMetadata</tt> implicitly.
 */
@Test
public void testCaseInsensitiveSchema() {
    ResultSetLoader rsLoader = new ResultSetLoaderImpl(fixture.allocator());
    RowSetLoader rootWriter = rsLoader.writer();
    TupleMetadata schema = rootWriter.tupleSchema();
    assertEquals(0, rsLoader.schemaVersion());
    // No columns defined in schema
    assertNull(schema.metadata("a"));
    try {
        schema.column(0);
        fail();
    } catch (IndexOutOfBoundsException e) {
    // Expected
    }
    try {
        rootWriter.column("a");
        fail();
    } catch (UndefinedColumnException e) {
    // Expected
    }
    try {
        rootWriter.column(0);
        fail();
    } catch (IndexOutOfBoundsException e) {
    // Expected
    }
    // Define a column
    assertEquals(0, rsLoader.schemaVersion());
    MaterializedField colSchema = SchemaBuilder.columnSchema("a", MinorType.VARCHAR, DataMode.REQUIRED);
    rootWriter.addColumn(colSchema);
    assertEquals(1, rsLoader.schemaVersion());
    // Can now be found, case insensitive
    assertTrue(colSchema.isEquivalent(schema.column(0)));
    ColumnMetadata colMetadata = schema.metadata(0);
    assertSame(colMetadata, schema.metadata("a"));
    assertSame(colMetadata, schema.metadata("A"));
    assertNotNull(rootWriter.column(0));
    assertNotNull(rootWriter.column("a"));
    assertNotNull(rootWriter.column("A"));
    assertEquals(1, schema.size());
    assertEquals(0, schema.index("a"));
    assertEquals(0, schema.index("A"));
    try {
        rootWriter.addColumn(colSchema);
        fail();
    } catch (UserException e) {
    // Expected
    }
    try {
        MaterializedField testCol = SchemaBuilder.columnSchema("A", MinorType.VARCHAR, DataMode.REQUIRED);
        rootWriter.addColumn(testCol);
        fail();
    } catch (UserException e) {
        // Expected
        assertTrue(e.getMessage().contains("Duplicate"));
    }
    // Can still add required fields while writing the first row.
    rsLoader.startBatch();
    rootWriter.start();
    rootWriter.scalar(0).setString("foo");
    MaterializedField col2 = SchemaBuilder.columnSchema("b", MinorType.VARCHAR, DataMode.REQUIRED);
    rootWriter.addColumn(col2);
    assertEquals(2, rsLoader.schemaVersion());
    assertTrue(col2.isEquivalent(schema.column(1)));
    ColumnMetadata col2Metadata = schema.metadata(1);
    assertSame(col2Metadata, schema.metadata("b"));
    assertSame(col2Metadata, schema.metadata("B"));
    assertEquals(2, schema.size());
    assertEquals(1, schema.index("b"));
    assertEquals(1, schema.index("B"));
    rootWriter.scalar(1).setString("second");
    // After first row, can add an optional or repeated.
    // Also allows a required field: values will be back-filled.
    rootWriter.save();
    rootWriter.start();
    rootWriter.scalar(0).setString("bar");
    rootWriter.scalar(1).setString("");
    MaterializedField col3 = SchemaBuilder.columnSchema("c", MinorType.VARCHAR, DataMode.REQUIRED);
    rootWriter.addColumn(col3);
    assertEquals(3, rsLoader.schemaVersion());
    assertTrue(col3.isEquivalent(schema.column(2)));
    ColumnMetadata col3Metadata = schema.metadata(2);
    assertSame(col3Metadata, schema.metadata("c"));
    assertSame(col3Metadata, schema.metadata("C"));
    assertEquals(3, schema.size());
    assertEquals(2, schema.index("c"));
    assertEquals(2, schema.index("C"));
    rootWriter.scalar("c").setString("c.2");
    MaterializedField col4 = SchemaBuilder.columnSchema("d", MinorType.VARCHAR, DataMode.OPTIONAL);
    rootWriter.addColumn(col4);
    assertEquals(4, rsLoader.schemaVersion());
    assertTrue(col4.isEquivalent(schema.column(3)));
    ColumnMetadata col4Metadata = schema.metadata(3);
    assertSame(col4Metadata, schema.metadata("d"));
    assertSame(col4Metadata, schema.metadata("D"));
    assertEquals(4, schema.size());
    assertEquals(3, schema.index("d"));
    assertEquals(3, schema.index("D"));
    rootWriter.scalar("d").setString("d.2");
    MaterializedField col5 = SchemaBuilder.columnSchema("e", MinorType.VARCHAR, DataMode.REPEATED);
    rootWriter.addColumn(col5);
    assertEquals(5, rsLoader.schemaVersion());
    assertTrue(col5.isEquivalent(schema.column(4)));
    ColumnMetadata col5Metadata = schema.metadata(4);
    assertSame(col5Metadata, schema.metadata("e"));
    assertSame(col5Metadata, schema.metadata("E"));
    assertEquals(5, schema.size());
    assertEquals(4, schema.index("e"));
    assertEquals(4, schema.index("E"));
    rootWriter.array(4).setObject(strArray("e1", "e2", "e3"));
    MaterializedField col6 = SchemaBuilder.columnSchema("f", MinorType.BIGINT, DataMode.REPEATED);
    rootWriter.addColumn(col6);
    assertEquals(6, rsLoader.schemaVersion());
    assertTrue(col6.isEquivalent(schema.column(5)));
    ColumnMetadata col6Metadata = schema.metadata(5);
    assertSame(col6Metadata, schema.metadata("f"));
    assertSame(col6Metadata, schema.metadata("F"));
    assertEquals(6, schema.size());
    assertEquals(5, schema.index("f"));
    assertEquals(5, schema.index("F"));
    rootWriter.array(5).setObject(new Long[] { Long.MIN_VALUE, Long.MAX_VALUE });
    MaterializedField col7 = SchemaBuilder.columnSchema("g", MinorType.INT, DataMode.REPEATED);
    rootWriter.addColumn(col7);
    assertEquals(7, rsLoader.schemaVersion());
    assertTrue(col7.isEquivalent(schema.column(6)));
    ColumnMetadata col7Metadata = schema.metadata(6);
    assertSame(col7Metadata, schema.metadata("g"));
    assertSame(col7Metadata, schema.metadata("G"));
    assertEquals(7, schema.size());
    assertEquals(6, schema.index("g"));
    assertEquals(6, schema.index("G"));
    rootWriter.array(6).setObject(new Integer[] { Integer.MIN_VALUE, Integer.MAX_VALUE });
    MaterializedField col8 = SchemaBuilder.columnSchema("h", MinorType.INT, DataMode.REPEATED);
    rootWriter.addColumn(col8);
    assertEquals(8, rsLoader.schemaVersion());
    assertTrue(col8.isEquivalent(schema.column(7)));
    ColumnMetadata col8Metadata = schema.metadata(7);
    assertSame(col8Metadata, schema.metadata("h"));
    assertSame(col8Metadata, schema.metadata("H"));
    assertEquals(8, schema.size());
    assertEquals(7, schema.index("h"));
    assertEquals(7, schema.index("H"));
    rootWriter.array(7).setObject(new Short[] { Short.MIN_VALUE, Short.MAX_VALUE });
    MaterializedField col9 = SchemaBuilder.columnSchema("i", MinorType.INT, DataMode.REPEATED);
    rootWriter.addColumn(col9);
    assertEquals(9, rsLoader.schemaVersion());
    assertTrue(col9.isEquivalent(schema.column(8)));
    ColumnMetadata col9Metadata = schema.metadata(8);
    assertSame(col9Metadata, schema.metadata("i"));
    assertSame(col9Metadata, schema.metadata("I"));
    assertEquals(9, schema.size());
    assertEquals(8, schema.index("i"));
    assertEquals(8, schema.index("I"));
    rootWriter.array(8).setObject(new Byte[] { Byte.MIN_VALUE, Byte.MAX_VALUE });
    MaterializedField col10 = SchemaBuilder.columnSchema("j", MinorType.FLOAT8, DataMode.REPEATED);
    rootWriter.addColumn(col10);
    assertEquals(10, rsLoader.schemaVersion());
    assertTrue(col10.isEquivalent(schema.column(9)));
    ColumnMetadata col10Metadata = schema.metadata(9);
    assertSame(col10Metadata, schema.metadata("j"));
    assertSame(col10Metadata, schema.metadata("J"));
    assertEquals(10, schema.size());
    assertEquals(9, schema.index("j"));
    assertEquals(9, schema.index("J"));
    rootWriter.array(9).setObject(new Double[] { Double.MIN_VALUE, Double.MAX_VALUE });
    MaterializedField col11 = SchemaBuilder.columnSchema("k", MinorType.FLOAT4, DataMode.REPEATED);
    rootWriter.addColumn(col11);
    assertEquals(11, rsLoader.schemaVersion());
    assertTrue(col11.isEquivalent(schema.column(10)));
    ColumnMetadata col11Metadata = schema.metadata(10);
    assertSame(col11Metadata, schema.metadata("k"));
    assertSame(col11Metadata, schema.metadata("K"));
    assertEquals(11, schema.size());
    assertEquals(10, schema.index("k"));
    assertEquals(10, schema.index("K"));
    rootWriter.array(10).setObject(new Float[] { Float.MIN_VALUE, Float.MAX_VALUE });
    MaterializedField col12 = SchemaBuilder.columnSchema("l", MinorType.BIT, DataMode.REPEATED);
    rootWriter.addColumn(col12);
    assertEquals(12, rsLoader.schemaVersion());
    assertTrue(col12.isEquivalent(schema.column(11)));
    ColumnMetadata col12Metadata = schema.metadata(11);
    assertSame(col12Metadata, schema.metadata("l"));
    assertSame(col12Metadata, schema.metadata("L"));
    assertEquals(12, schema.size());
    assertEquals(11, schema.index("l"));
    assertEquals(11, schema.index("L"));
    rootWriter.array(11).setObject(new Boolean[] { Boolean.TRUE, Boolean.FALSE });
    rootWriter.save();
    // Verify. No reason to expect problems, but might as well check.
    RowSet result = fixture.wrap(rsLoader.harvest());
    assertEquals(12, rsLoader.schemaVersion());
    SingleRowSet expected = fixture.rowSetBuilder(result.batchSchema()).addRow("foo", "second", "", null, strArray(), longArray(), intArray(), shortArray(), byteArray(), doubleArray(), floatArray(), boolArray()).addRow("bar", "", "c.2", "d.2", strArray("e1", "e2", "e3"), longArray(Long.MIN_VALUE, Long.MAX_VALUE), intArray(Integer.MIN_VALUE, Integer.MAX_VALUE), shortArray(Short.MIN_VALUE, Short.MAX_VALUE), byteArray((int) Byte.MIN_VALUE, (int) Byte.MAX_VALUE), doubleArray(Double.MIN_VALUE, Double.MAX_VALUE), floatArray(Float.MIN_VALUE, Float.MAX_VALUE), boolArray(Boolean.TRUE, Boolean.FALSE)).build();
    RowSetUtilities.verify(expected, result);
    // Handy way to test that close works to abort an in-flight batch
    // and clean up.
    rsLoader.close();
}
Also used : ColumnMetadata(org.apache.drill.exec.record.metadata.ColumnMetadata) 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) SingleRowSet(org.apache.drill.exec.physical.rowSet.RowSet.SingleRowSet) RowSet(org.apache.drill.exec.physical.rowSet.RowSet) MaterializedField(org.apache.drill.exec.record.MaterializedField) UserException(org.apache.drill.common.exceptions.UserException) RowSetLoader(org.apache.drill.exec.physical.resultSet.RowSetLoader) UndefinedColumnException(org.apache.drill.exec.vector.accessor.TupleWriter.UndefinedColumnException) SubOperatorTest(org.apache.drill.test.SubOperatorTest) Test(org.junit.Test)

Example 77 with ColumnMetadata

use of org.apache.drill.exec.record.metadata.ColumnMetadata in project drill by apache.

the class TestSchemaParser method checkSchema.

private void checkSchema(String schemaString, TupleMetadata expectedSchema) throws IOException {
    TupleMetadata actualSchema = SchemaExprParser.parseSchema(schemaString);
    assertEquals(expectedSchema.size(), actualSchema.size());
    assertEquals(expectedSchema.properties(), actualSchema.properties());
    expectedSchema.toMetadataList().forEach(expectedMetadata -> {
        ColumnMetadata actualMetadata = actualSchema.metadata(expectedMetadata.name());
        assertEquals(expectedMetadata.columnString(), actualMetadata.columnString());
    });
}
Also used : DictColumnMetadata(org.apache.drill.exec.record.metadata.DictColumnMetadata) ColumnMetadata(org.apache.drill.exec.record.metadata.ColumnMetadata) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata)

Example 78 with ColumnMetadata

use of org.apache.drill.exec.record.metadata.ColumnMetadata in project drill by apache.

the class TestSchemaParser method testFormat.

@Test
public void testFormat() throws Exception {
    String value = "`a` DATE NOT NULL FORMAT 'yyyy-MM-dd'";
    TupleMetadata schema = SchemaExprParser.parseSchema(value);
    ColumnMetadata columnMetadata = schema.metadata("a");
    assertEquals("yyyy-MM-dd", columnMetadata.format());
    assertEquals(value, columnMetadata.columnString());
}
Also used : DictColumnMetadata(org.apache.drill.exec.record.metadata.DictColumnMetadata) ColumnMetadata(org.apache.drill.exec.record.metadata.ColumnMetadata) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) Test(org.junit.Test) BaseTest(org.apache.drill.test.BaseTest)

Example 79 with ColumnMetadata

use of org.apache.drill.exec.record.metadata.ColumnMetadata in project drill by apache.

the class TestSchemaParser method testParseColumn.

@Test
public void testParseColumn() throws Exception {
    ColumnMetadata column = SchemaExprParser.parseColumn("col int not null");
    assertEquals("`col` INT NOT NULL", column.columnString());
}
Also used : DictColumnMetadata(org.apache.drill.exec.record.metadata.DictColumnMetadata) ColumnMetadata(org.apache.drill.exec.record.metadata.ColumnMetadata) Test(org.junit.Test) BaseTest(org.apache.drill.test.BaseTest)

Example 80 with ColumnMetadata

use of org.apache.drill.exec.record.metadata.ColumnMetadata in project drill by apache.

the class TestSchemaParser method testColumnProperties.

@Test
public void testColumnProperties() throws Exception {
    String value = "`a` INT NOT NULL PROPERTIES { 'k1' = 'v1', 'k2' = 'v2' }";
    TupleMetadata schema = SchemaExprParser.parseSchema(value);
    ColumnMetadata columnMetadata = schema.metadata("a");
    Map<String, String> properties = new LinkedHashMap<>();
    properties.put("k1", "v1");
    properties.put("k2", "v2");
    assertEquals(properties, columnMetadata.properties());
    assertEquals(value, columnMetadata.columnString());
}
Also used : DictColumnMetadata(org.apache.drill.exec.record.metadata.DictColumnMetadata) ColumnMetadata(org.apache.drill.exec.record.metadata.ColumnMetadata) TupleMetadata(org.apache.drill.exec.record.metadata.TupleMetadata) LinkedHashMap(java.util.LinkedHashMap) Test(org.junit.Test) BaseTest(org.apache.drill.test.BaseTest)

Aggregations

ColumnMetadata (org.apache.drill.exec.record.metadata.ColumnMetadata)195 Test (org.junit.Test)104 TupleMetadata (org.apache.drill.exec.record.metadata.TupleMetadata)98 SchemaBuilder (org.apache.drill.exec.record.metadata.SchemaBuilder)50 DrillTest (org.apache.drill.test.DrillTest)37 PrimitiveColumnMetadata (org.apache.drill.exec.record.metadata.PrimitiveColumnMetadata)31 SubOperatorTest (org.apache.drill.test.SubOperatorTest)31 BaseTest (org.apache.drill.test.BaseTest)26 MapColumnMetadata (org.apache.drill.exec.record.metadata.MapColumnMetadata)20 SchemaBuilder (org.apache.drill.test.rowSet.schema.SchemaBuilder)20 VariantColumnMetadata (org.apache.drill.exec.record.metadata.VariantColumnMetadata)19 VariantMetadata (org.apache.drill.exec.record.metadata.VariantMetadata)19 AbstractColumnMetadata (org.apache.drill.exec.record.metadata.AbstractColumnMetadata)17 ScalarWriter (org.apache.drill.exec.vector.accessor.ScalarWriter)17 DictColumnMetadata (org.apache.drill.exec.record.metadata.DictColumnMetadata)15 EvfTest (org.apache.drill.categories.EvfTest)13 ProjectionFilter (org.apache.drill.exec.physical.resultSet.impl.ProjectionFilter)13 TupleSchema (org.apache.drill.exec.record.metadata.TupleSchema)11 ArrayList (java.util.ArrayList)10 ProjResult (org.apache.drill.exec.physical.resultSet.impl.ProjectionFilter.ProjResult)10