Examples with FileMetadataColumn org.apache.drill.exec.physical.impl.scan.file.FileMetadataColumn used on opensource projects

Example 1 with FileMetadataColumn

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

the class ReaderSchemaOrchestrator method projectMetadata.

/**
 * Sets {@code PROJECT_METADATA} implicit column value.
 *
 * @param projectMetadata whether {@code PROJECT_METADATA} should be replaced
 * @return {@code true} if {@code PROJECT_METADATA} implicit column
 * is present in scan projection and its value is updated
 */
private boolean projectMetadata(boolean projectMetadata) {
    ImplicitInternalFileColumns original;
    ImplicitInternalFileColumns newColumn;
    if (projectMetadata) {
        original = ImplicitInternalFileColumns.USE_METADATA;
        newColumn = ImplicitInternalFileColumns.PROJECT_METADATA;
    } else {
        original = ImplicitInternalFileColumns.PROJECT_METADATA;
        newColumn = ImplicitInternalFileColumns.USE_METADATA;
    }
    List<ColumnProjection> outputColumns = scanOrchestrator.scanProj.columns();
    for (int i = 0; i < outputColumns.size(); i++) {
        ColumnProjection outputColumn = outputColumns.get(i);
        if (outputColumn instanceof FileMetadataColumn) {
            FileMetadataColumn metadataColumn = (FileMetadataColumn) outputColumn;
            if (metadataColumn.defn().defn == original) {
                projectMetadata = scanOrchestrator.scanProj.requestedCols().stream().anyMatch(SchemaPath.getSimplePath(metadataColumn.name())::equals);
                if (projectMetadata) {
                    outputColumns.set(i, new FileMetadataColumn(metadataColumn.name(), new FileMetadataColumnDefn(metadataColumn.defn().colName(), newColumn)));
                }
                return projectMetadata;
            }
        }
    }
    return false;
}

Example 2 with FileMetadataColumn

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

the class TestImplicitColumnProjection method testProjectList.

/**
 * Test the file projection planner with metadata.
 */
@Test
public void testProjectList() {
    Path filePath = new Path("hdfs:///w/x/y/z.csv");
    ImplicitColumnManager metadataManager = new ImplicitColumnManager(fixture.getOptionManager(), standardOptions(filePath));
    ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList(ScanTestUtils.FILE_NAME_COL, "a", ScanTestUtils.partitionColName(0)), ScanTestUtils.parsers(metadataManager.projectionParser()));
    assertEquals(3, scanProj.columns().size());
    // Scan-level projection: defines the columns
    {
        assertTrue(scanProj.columns().get(0) instanceof FileMetadataColumn);
        FileMetadataColumn col0 = (FileMetadataColumn) scanProj.columns().get(0);
        assertTrue(col0 instanceof FileMetadataColumn);
        assertEquals(ScanTestUtils.FILE_NAME_COL, col0.name());
        assertEquals(MinorType.VARCHAR, col0.schema().getType().getMinorType());
        assertEquals(DataMode.REQUIRED, col0.schema().getType().getMode());
        ColumnProjection col1 = scanProj.columns().get(1);
        assertTrue(col1 instanceof UnresolvedColumn);
        assertEquals("a", col1.name());
        assertTrue(scanProj.columns().get(2) instanceof PartitionColumn);
        PartitionColumn col2 = (PartitionColumn) scanProj.columns().get(2);
        assertTrue(col2 instanceof PartitionColumn);
        assertEquals(ScanTestUtils.partitionColName(0), col2.name());
        assertEquals(MinorType.VARCHAR, col2.schema().getType().getMinorType());
        assertEquals(DataMode.OPTIONAL, col2.schema().getType().getMode());
    }
    // Schema-level projection, fills in values.
    TupleMetadata tableSchema = new SchemaBuilder().add("a", MinorType.VARCHAR).buildSchema();
    metadataManager.startFile(filePath);
    NullColumnBuilder builder = new NullBuilderBuilder().build();
    ResolvedRow rootTuple = new ResolvedRow(builder);
    new ExplicitSchemaProjection(scanProj, tableSchema, rootTuple, ScanTestUtils.resolvers(metadataManager));
    List<ResolvedColumn> columns = rootTuple.columns();
    assertEquals(3, columns.size());
    {
        assertTrue(columns.get(0) instanceof FileMetadataColumn);
        FileMetadataColumn col0 = (FileMetadataColumn) columns.get(0);
        assertTrue(col0 instanceof FileMetadataColumn);
        assertEquals(ScanTestUtils.FILE_NAME_COL, col0.name());
        assertEquals("z.csv", col0.value());
        assertEquals(MinorType.VARCHAR, col0.schema().getType().getMinorType());
        assertEquals(DataMode.REQUIRED, col0.schema().getType().getMode());
        ResolvedColumn col1 = columns.get(1);
        assertEquals("a", col1.name());
        assertTrue(columns.get(2) instanceof PartitionColumn);
        PartitionColumn col2 = (PartitionColumn) columns.get(2);
        assertTrue(col2 instanceof PartitionColumn);
        assertEquals(ScanTestUtils.partitionColName(0), col2.name());
        assertEquals("x", col2.value());
        assertEquals(MinorType.VARCHAR, col2.schema().getType().getMinorType());
        assertEquals(DataMode.OPTIONAL, col2.schema().getType().getMode());
    }
    // Verify that the file metadata columns were picked out
    assertEquals(2, metadataManager.metadataColumns().size());
    assertSame(columns.get(0), metadataManager.metadataColumns().get(0));
    assertSame(columns.get(2), metadataManager.metadataColumns().get(1));
}

Example 3 with FileMetadataColumn

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

the class TestColumnsArrayParser method testMetadataColumnsWithColumnsArray.

/**
 * The `columns` column is special: can't be used with other column names.
 * Make sure that the rule <i>does not</i> apply to implicit columns.
 */
@Test
public void testMetadataColumnsWithColumnsArray() {
    Path filePath = new Path("hdfs:///w/x/y/z.csv");
    ImplicitColumnManager metadataManager = new ImplicitColumnManager(fixture.getOptionManager(), standardOptions(filePath));
    ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList(ScanTestUtils.FILE_NAME_COL, ColumnsScanFramework.COLUMNS_COL, ScanTestUtils.SUFFIX_COL), ScanTestUtils.parsers(new ColumnsArrayParser(true), metadataManager.projectionParser()));
    assertFalse(scanProj.projectAll());
    assertEquals(3, scanProj.columns().size());
    assertEquals(ScanTestUtils.FILE_NAME_COL, scanProj.columns().get(0).name());
    assertEquals(ColumnsScanFramework.COLUMNS_COL, scanProj.columns().get(1).name());
    assertEquals(ScanTestUtils.SUFFIX_COL, scanProj.columns().get(2).name());
    // Verify column type
    assertTrue(scanProj.columns().get(0) instanceof FileMetadataColumn);
    assertTrue(scanProj.columns().get(1) instanceof UnresolvedColumnsArrayColumn);
    assertTrue(scanProj.columns().get(2) instanceof FileMetadataColumn);
}

Example 4 with FileMetadataColumn

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

the class TestImplicitColumnParser method testFileMetadataColumnSelection.

/**
 * Test including file metadata (AKA "implicit columns") in the project
 * list.
 */
@Test
public void testFileMetadataColumnSelection() {
    Path filePath = new Path("hdfs:///w/x/y/z.csv");
    ImplicitColumnManager implictColManager = new ImplicitColumnManager(fixture.getOptionManager(), standardOptions(filePath));
    // Simulate SELECT a, fqn, filEPath, filename, suffix ...
    ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList("a", ScanTestUtils.FULLY_QUALIFIED_NAME_COL, // Sic, to test case sensitivity
    "filEPath", ScanTestUtils.FILE_NAME_COL, ScanTestUtils.SUFFIX_COL), Lists.newArrayList(implictColManager.projectionParser()));
    assertFalse(scanProj.projectAll());
    assertEquals(5, scanProj.requestedCols().size());
    assertEquals(5, scanProj.columns().size());
    assertEquals("a", scanProj.columns().get(0).name());
    assertEquals(ScanTestUtils.FULLY_QUALIFIED_NAME_COL, scanProj.columns().get(1).name());
    assertEquals("filEPath", scanProj.columns().get(2).name());
    assertEquals(ScanTestUtils.FILE_NAME_COL, scanProj.columns().get(3).name());
    assertEquals(ScanTestUtils.SUFFIX_COL, scanProj.columns().get(4).name());
    // Verify column type
    assertTrue(scanProj.columns().get(0) instanceof UnresolvedColumn);
    assertTrue(scanProj.columns().get(1) instanceof FileMetadataColumn);
    assertTrue(scanProj.columns().get(2) instanceof FileMetadataColumn);
    assertTrue(scanProj.columns().get(3) instanceof FileMetadataColumn);
    assertTrue(scanProj.columns().get(4) instanceof FileMetadataColumn);
    assertTrue(implictColManager.hasImplicitCols());
}

Example 5 with FileMetadataColumn

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

the class TestImplicitColumnParser method testShadowed.

/**
 * Verify that names that look like metadata columns, but appear
 * to be maps or arrays, are not interpreted as metadata. That is,
 * the projected table map or array "shadows" the metadata column.
 */
@Test
public void testShadowed() {
    Path filePath = new Path("hdfs:///w/x/y/z.csv");
    ImplicitColumnManager implictColManager = new ImplicitColumnManager(fixture.getOptionManager(), standardOptions(filePath));
    ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList(ScanTestUtils.FILE_NAME_COL + ".a", ScanTestUtils.FILE_PATH_COL + "[0]", ScanTestUtils.partitionColName(0) + ".b", ScanTestUtils.partitionColName(1) + "[0]", ScanTestUtils.SUFFIX_COL), Lists.newArrayList(implictColManager.projectionParser()));
    List<ColumnProjection> cols = scanProj.columns();
    assertEquals(5, cols.size());
    for (int i = 0; i < 4; i++) {
        assertTrue(scanProj.columns().get(i) instanceof UnresolvedColumn);
    }
    assertTrue(scanProj.columns().get(4) instanceof FileMetadataColumn);
}