marton-bod commented on code in PR #3131:
URL: https://github.com/apache/hive/pull/3131#discussion_r845988600
##########
iceberg/iceberg-handler/src/test/java/org/apache/iceberg/mr/hive/TestHiveIcebergV2.java:
##########
@@ -228,6 +230,104 @@ public void
testReadAndWriteFormatV2Partitioned_PosDelete_RowSupplied() throws I
Assert.assertArrayEquals(new Object[] {2L, "Trudy", "Pink"},
objects.get(3));
}
+ @Test
+ public void testDeleteStatementUnpartitioned() {
+ Assume.assumeFalse("Iceberg DELETEs are only implemented for
non-vectorized mode for now", isVectorized);
+
+ // create and insert an initial batch of records
+ testTables.createTable(shell, "customers",
HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
+ PartitionSpec.unpartitioned(), fileFormat,
HiveIcebergStorageHandlerTestUtils.OTHER_CUSTOMER_RECORDS_2, 2);
+ // insert one more batch so that we have multiple data files within the
same partition
+
shell.executeStatement(testTables.getInsertQuery(HiveIcebergStorageHandlerTestUtils.OTHER_CUSTOMER_RECORDS_1,
+ TableIdentifier.of("default", "customers"), false));
+
+ shell.executeStatement("DELETE FROM customers WHERE customer_id=3 or
first_name='Joanna'");
+
+ List<Object[]> objects = shell.executeStatement("SELECT * FROM customers
ORDER BY customer_id, last_name");
+ Assert.assertEquals(6, objects.size());
+ List<Record> expected =
TestHelper.RecordsBuilder.newInstance(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA)
+ .add(1L, "Sharon", "Taylor")
+ .add(2L, "Jake", "Donnel")
+ .add(2L, "Susan", "Morrison")
+ .add(2L, "Bob", "Silver")
+ .add(4L, "Laci", "Zold")
+ .add(5L, "Peti", "Rozsaszin")
+ .build();
+ HiveIcebergTestUtils.validateData(expected,
+
HiveIcebergTestUtils.valueForRow(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
objects), 0);
+ }
+
+ @Test
+ public void testDeleteStatementPartitioned() {
+ Assume.assumeFalse("Iceberg DELETEs are only implemented for
non-vectorized mode for now", isVectorized);
+ PartitionSpec spec =
PartitionSpec.builderFor(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA)
+ .identity("last_name").bucket("customer_id", 16).build();
+
+ // create and insert an initial batch of records
+ testTables.createTable(shell, "customers",
HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
+ spec, fileFormat,
HiveIcebergStorageHandlerTestUtils.OTHER_CUSTOMER_RECORDS_2, 2);
+ // insert one more batch so that we have multiple data files within the
same partition
+
shell.executeStatement(testTables.getInsertQuery(HiveIcebergStorageHandlerTestUtils.OTHER_CUSTOMER_RECORDS_1,
+ TableIdentifier.of("default", "customers"), false));
+
+ shell.executeStatement("DELETE FROM customers WHERE customer_id=3 or
first_name='Joanna'");
+
+ List<Object[]> objects = shell.executeStatement("SELECT * FROM customers
ORDER BY customer_id, last_name");
+ Assert.assertEquals(6, objects.size());
+ List<Record> expected =
TestHelper.RecordsBuilder.newInstance(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA)
+ .add(1L, "Sharon", "Taylor")
+ .add(2L, "Jake", "Donnel")
+ .add(2L, "Susan", "Morrison")
+ .add(2L, "Bob", "Silver")
+ .add(4L, "Laci", "Zold")
+ .add(5L, "Peti", "Rozsaszin")
+ .build();
+ HiveIcebergTestUtils.validateData(expected,
+
HiveIcebergTestUtils.valueForRow(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
objects), 0);
+ }
+
+ @Test
+ public void testDeleteStatementWithOtherTable() {
+ Assume.assumeFalse("Iceberg DELETEs are only implemented for
non-vectorized mode for now", isVectorized);
+ PartitionSpec spec =
PartitionSpec.builderFor(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA)
+ .identity("last_name").bucket("customer_id", 16).build();
+
+ // create a couple of tables, with an initial batch of records
+ testTables.createTable(shell, "customers",
HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
+ spec, fileFormat,
HiveIcebergStorageHandlerTestUtils.OTHER_CUSTOMER_RECORDS_2, 2);
+ testTables.createTable(shell, "other",
HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
+ spec, fileFormat,
HiveIcebergStorageHandlerTestUtils.OTHER_CUSTOMER_RECORDS_1, 2);
+
+ shell.executeStatement("DELETE FROM customers WHERE customer_id in (select
t1.customer_id from customers t1 join " +
+ "other t2 on t1.customer_id = t2.customer_id) or " +
+ "first_name in (select first_name from customers where first_name =
'Bob')");
+
+ List<Object[]> objects = shell.executeStatement("SELECT * FROM customers
ORDER BY customer_id, last_name");
+ Assert.assertEquals(5, objects.size());
+ List<Record> expected =
TestHelper.RecordsBuilder.newInstance(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA)
+ .add(1L, "Joanna", "Pierce")
+ .add(1L, "Sharon", "Taylor")
+ .add(2L, "Jake", "Donnel")
+ .add(2L, "Susan", "Morrison")
+ .add(2L, "Joanna", "Silver")
+ .build();
+ HiveIcebergTestUtils.validateData(expected,
+
HiveIcebergTestUtils.valueForRow(HiveIcebergStorageHandlerTestUtils.CUSTOMER_SCHEMA,
objects), 0);
+ }
+
+ @Test
+ public void testDeleteStatementThrowsIfVectorizationEnabled() {
Review Comment:
Actually, the vectorized reader currently can't handle delete files, so we
can give back incorrect results for v2 tables when vectorization is on. So I've
added logic to fallback to non-vector mode for V2 tables for now. On the
positive side, that means that the user does not need to set vector=off manually
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
