use of org.apache.drill.exec.physical.impl.scan.project.ScanLevelProjection in project drill by apache.
the class TestImplicitColumnProjection method testPartitionColumnTwoDigits.
/**
* Test the obscure case that the partition column contains two digits:
* dir11. Also tests the obscure case that the output only has partition
* columns.
*/
@Test
public void testPartitionColumnTwoDigits() {
Path filePath = new Path("hdfs:///x/0/1/2/3/4/5/6/7/8/9/10/d11/z.csv");
ImplicitColumnManager metadataManager = new ImplicitColumnManager(fixture.getOptionManager(), standardOptions(filePath));
ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList("dir11"), ScanTestUtils.parsers(metadataManager.projectionParser()));
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(1, columns.size());
assertEquals("d11", ((MetadataColumn) columns.get(0)).value());
}
use of org.apache.drill.exec.physical.impl.scan.project.ScanLevelProjection 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));
}
use of org.apache.drill.exec.physical.impl.scan.project.ScanLevelProjection in project drill by apache.
the class TestColumnsArrayParser method testColumnsElements.
@Test
public void testColumnsElements() {
ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList(ColumnsScanFramework.COLUMNS_COL + "[3]", ColumnsScanFramework.COLUMNS_COL + "[1]"), ScanTestUtils.parsers(new ColumnsArrayParser(true)));
assertFalse(scanProj.projectAll());
assertEquals(2, scanProj.requestedCols().size());
assertEquals(1, scanProj.columns().size());
assertEquals(ColumnsScanFramework.COLUMNS_COL, scanProj.columns().get(0).name());
// Verify column type
assertTrue(scanProj.columns().get(0) instanceof UnresolvedColumnsArrayColumn);
UnresolvedColumnsArrayColumn colsCol = (UnresolvedColumnsArrayColumn) scanProj.columns().get(0);
boolean[] indexes = colsCol.selectedIndexes();
assertNotNull(indexes);
assertEquals(4, indexes.length);
assertFalse(indexes[0]);
assertTrue(indexes[1]);
assertFalse(indexes[0]);
assertTrue(indexes[1]);
}
use of org.apache.drill.exec.physical.impl.scan.project.ScanLevelProjection 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);
}
use of org.apache.drill.exec.physical.impl.scan.project.ScanLevelProjection in project drill by apache.
the class TestColumnsArrayParser method testTwoColumnsArray.
/**
* Requesting `columns` twice: second is ignored.
*/
@Test
public void testTwoColumnsArray() {
ScanLevelProjection scanProj = ScanLevelProjection.build(RowSetTestUtils.projectList(ColumnsScanFramework.COLUMNS_COL, ColumnsScanFramework.COLUMNS_COL), ScanTestUtils.parsers(new ColumnsArrayParser(false)));
assertFalse(scanProj.projectAll());
assertEquals(2, scanProj.requestedCols().size());
assertEquals(1, scanProj.columns().size());
assertEquals(ColumnsScanFramework.COLUMNS_COL, scanProj.columns().get(0).name());
}
Aggregations