use of org.apache.iceberg.deletes.PositionDelete in project hive by apache.
the class TestHiveIcebergV2 method testReadAndWriteFormatV2Partitioned_PosDelete_RowSupplied.
@Test
public void testReadAndWriteFormatV2Partitioned_PosDelete_RowSupplied() throws IOException {
Assume.assumeFalse("Reading V2 tables with delete files are only supported currently in " + "non-vectorized mode and only Parquet/Avro", isVectorized || fileFormat == FileFormat.ORC);
PartitionSpec spec = PartitionSpec.builderFor(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA).identity("customer_id").build();
Table tbl = testTables.createTable(shell, "customers", HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA, spec, fileFormat, HiveIcebergStorageHandlerTestUtils.CUSTOMER_RECORDS, 2);
// add some more data to the same partition
shell.executeStatement("insert into customers values (0, 'Laura', 'Yellow'), (0, 'John', 'Green'), " + "(0, 'Blake', 'Blue')");
tbl.refresh();
// delete the first and third rows from the newly-added data file
DataFile dataFile = StreamSupport.stream(tbl.currentSnapshot().addedFiles().spliterator(), false).filter(file -> file.partition().get(0, Long.class) == 0L).filter(file -> file.recordCount() == 3).findAny().orElseThrow(() -> new RuntimeException("Did not find the desired data file in the test table"));
List<Record> rowsToDel = TestHelper.RecordsBuilder.newInstance(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA).add(0L, "Laura", "Yellow").add(0L, "Blake", "Blue").build();
List<PositionDelete<Record>> deletes = ImmutableList.of(positionDelete(dataFile.path(), 0L, rowsToDel.get(0)), positionDelete(dataFile.path(), 2L, rowsToDel.get(1)));
DeleteFile deleteFile = HiveIcebergTestUtils.createPositionalDeleteFile(tbl, "dummyPath", fileFormat, ImmutableMap.of("customer_id", 0L), deletes);
tbl.newRowDelta().addDeletes(deleteFile).commit();
List<Object[]> objects = shell.executeStatement("SELECT * FROM customers ORDER BY customer_id, first_name");
Assert.assertEquals(4, objects.size());
Assert.assertArrayEquals(new Object[] { 0L, "Alice", "Brown" }, objects.get(0));
Assert.assertArrayEquals(new Object[] { 0L, "John", "Green" }, objects.get(1));
Assert.assertArrayEquals(new Object[] { 1L, "Bob", "Green" }, objects.get(2));
Assert.assertArrayEquals(new Object[] { 2L, "Trudy", "Pink" }, objects.get(3));
}
use of org.apache.iceberg.deletes.PositionDelete in project hive by apache.
the class TestHiveIcebergV2 method testReadAndWriteFormatV2Unpartitioned_PosDelete.
@Test
public void testReadAndWriteFormatV2Unpartitioned_PosDelete() throws IOException {
Assume.assumeFalse("Reading V2 tables with delete files are only supported currently in " + "non-vectorized mode and only Parquet/Avro", isVectorized || fileFormat == FileFormat.ORC);
Table tbl = testTables.createTable(shell, "customers", HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA, PartitionSpec.unpartitioned(), fileFormat, HiveIcebergStorageHandlerTestUtils.CUSTOMER_RECORDS, 2);
// delete one of the rows
DataFile dataFile = StreamSupport.stream(tbl.currentSnapshot().addedFiles().spliterator(), false).findFirst().orElseThrow(() -> new RuntimeException("Did not find any data files for test table"));
List<PositionDelete<Record>> deletes = ImmutableList.of(positionDelete(dataFile.path(), 2L, HiveIcebergStorageHandlerTestUtils.CUSTOMER_RECORDS.get(2)));
DeleteFile deleteFile = HiveIcebergTestUtils.createPositionalDeleteFile(tbl, "dummyPath", fileFormat, null, deletes);
tbl.newRowDelta().addDeletes(deleteFile).commit();
List<Object[]> objects = shell.executeStatement("SELECT * FROM customers ORDER BY customer_id");
// only the other two rows are present
Assert.assertEquals(2, objects.size());
Assert.assertArrayEquals(new Object[] { 0L, "Alice", "Brown" }, objects.get(0));
Assert.assertArrayEquals(new Object[] { 1L, "Bob", "Green" }, objects.get(1));
}
use of org.apache.iceberg.deletes.PositionDelete in project hive by apache.
the class TestHiveIcebergV2 method testReadAndWriteFormatV2Partitioned_PosDelete_RowNotSupplied.
@Test
public void testReadAndWriteFormatV2Partitioned_PosDelete_RowNotSupplied() throws IOException {
Assume.assumeFalse("Reading V2 tables with delete files are only supported currently in " + "non-vectorized mode and only Parquet/Avro", isVectorized || fileFormat == FileFormat.ORC);
PartitionSpec spec = PartitionSpec.builderFor(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA).identity("customer_id").build();
Table tbl = testTables.createTable(shell, "customers", HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA, spec, fileFormat, HiveIcebergStorageHandlerTestUtils.CUSTOMER_RECORDS, 2);
// add some more data to the same partition
shell.executeStatement("insert into customers values (0, 'Laura', 'Yellow'), (0, 'John', 'Green'), " + "(0, 'Blake', 'Blue')");
tbl.refresh();
// delete the first and third rows from the newly-added data file - with row supplied
DataFile dataFile = StreamSupport.stream(tbl.currentSnapshot().addedFiles().spliterator(), false).filter(file -> file.partition().get(0, Long.class) == 0L).filter(file -> file.recordCount() == 3).findAny().orElseThrow(() -> new RuntimeException("Did not find the desired data file in the test table"));
List<PositionDelete<Record>> deletes = ImmutableList.of(positionDelete(dataFile.path(), 0L, null), positionDelete(dataFile.path(), 2L, null));
DeleteFile deleteFile = HiveIcebergTestUtils.createPositionalDeleteFile(tbl, "dummyPath", fileFormat, ImmutableMap.of("customer_id", 0L), deletes);
tbl.newRowDelta().addDeletes(deleteFile).commit();
List<Object[]> objects = shell.executeStatement("SELECT * FROM customers ORDER BY customer_id, first_name");
Assert.assertEquals(4, objects.size());
Assert.assertArrayEquals(new Object[] { 0L, "Alice", "Brown" }, objects.get(0));
Assert.assertArrayEquals(new Object[] { 0L, "John", "Green" }, objects.get(1));
Assert.assertArrayEquals(new Object[] { 1L, "Bob", "Green" }, objects.get(2));
Assert.assertArrayEquals(new Object[] { 2L, "Trudy", "Pink" }, objects.get(3));
}
Aggregations