Examples with GetSplitsResponse com.amazonaws.athena.connector.lambda.metadata.GetSplitsResponse used on opensource projects

Example 1 with GetSplitsResponse

use of com.amazonaws.athena.connector.lambda.metadata.GetSplitsResponse in project aws-athena-query-federation by awslabs.

the class CompositeHandlerTest method setUp.

@Before
public void setUp() throws Exception {
    logger.info("{}: enter", testName.getMethodName());
    allocator = new BlockAllocatorImpl();
    objectMapper = ObjectMapperFactory.create(allocator);
    mockMetadataHandler = mock(MetadataHandler.class);
    mockRecordHandler = mock(RecordHandler.class);
    schemaForRead = SchemaBuilder.newBuilder().addField("col1", new ArrowType.Int(32, true)).build();
    when(mockMetadataHandler.doGetTableLayout(any(BlockAllocatorImpl.class), any(GetTableLayoutRequest.class))).thenReturn(new GetTableLayoutResponse("catalog", new TableName("schema", "table"), BlockUtils.newBlock(allocator, "col1", Types.MinorType.BIGINT.getType(), 1L)));
    when(mockMetadataHandler.doListTables(any(BlockAllocatorImpl.class), any(ListTablesRequest.class))).thenReturn(new ListTablesResponse("catalog", Collections.singletonList(new TableName("schema", "table")), null));
    when(mockMetadataHandler.doGetTable(any(BlockAllocatorImpl.class), any(GetTableRequest.class))).thenReturn(new GetTableResponse("catalog", new TableName("schema", "table"), SchemaBuilder.newBuilder().addStringField("col1").build()));
    when(mockMetadataHandler.doListSchemaNames(any(BlockAllocatorImpl.class), any(ListSchemasRequest.class))).thenReturn(new ListSchemasResponse("catalog", Collections.singleton("schema1")));
    when(mockMetadataHandler.doGetSplits(any(BlockAllocatorImpl.class), any(GetSplitsRequest.class))).thenReturn(new GetSplitsResponse("catalog", Split.newBuilder(null, null).build()));
    when(mockMetadataHandler.doPing(any(PingRequest.class))).thenReturn(new PingResponse("catalog", "queryId", "type", 23, 2));
    when(mockRecordHandler.doReadRecords(any(BlockAllocatorImpl.class), any(ReadRecordsRequest.class))).thenReturn(new ReadRecordsResponse("catalog", BlockUtils.newEmptyBlock(allocator, "col", new ArrowType.Int(32, true))));
    compositeHandler = new CompositeHandler(mockMetadataHandler, mockRecordHandler);
}

Example 2 with GetSplitsResponse

use of com.amazonaws.athena.connector.lambda.metadata.GetSplitsResponse in project aws-athena-query-federation by awslabs.

the class GlueMetadataHandlerTest method setUp.

@Before
public void setUp() throws Exception {
    logger.info("{}: enter", testName.getMethodName());
    handler = new GlueMetadataHandler(mockGlue, new LocalKeyFactory(), mock(AWSSecretsManager.class), mock(AmazonAthena.class), "glue-test", "spill-bucket", "spill-prefix") {

        @Override
        public GetTableLayoutResponse doGetTableLayout(BlockAllocator blockAllocator, GetTableLayoutRequest request) {
            throw new UnsupportedOperationException();
        }

        @Override
        public void getPartitions(BlockWriter blockWriter, GetTableLayoutRequest request, QueryStatusChecker queryStatusChecker) throws Exception {
            throw new UnsupportedOperationException();
        }

        @Override
        public GetSplitsResponse doGetSplits(BlockAllocator blockAllocator, GetSplitsRequest request) {
            throw new UnsupportedOperationException();
        }
    };
    allocator = new BlockAllocatorImpl();
    // doListTables pagination.
    when(mockGlue.getTables(any(GetTablesRequest.class))).thenAnswer((InvocationOnMock invocationOnMock) -> {
        GetTablesRequest request = (GetTablesRequest) invocationOnMock.getArguments()[0];
        String nextToken = request.getNextToken();
        int pageSize = request.getMaxResults() == null ? UNLIMITED_PAGE_SIZE_VALUE : request.getMaxResults();
        assertEquals(accountId, request.getCatalogId());
        assertEquals(schema, request.getDatabaseName());
        GetTablesResult mockResult = mock(GetTablesResult.class);
        if (pageSize == UNLIMITED_PAGE_SIZE_VALUE) {
            // Simulate full list of tables returned from Glue.
            when(mockResult.getTableList()).thenReturn(unPaginatedTables);
            when(mockResult.getNextToken()).thenReturn(null);
        } else {
            // Simulate paginated list of tables returned from Glue.
            List<Table> paginatedTables = unPaginatedTables.stream().sorted(Comparator.comparing(Table::getName)).filter(table -> nextToken == null || table.getName().compareTo(nextToken) >= 0).limit(pageSize + 1).collect(Collectors.toList());
            if (paginatedTables.size() > pageSize) {
                when(mockResult.getNextToken()).thenReturn(paginatedTables.get(pageSize).getName());
                when(mockResult.getTableList()).thenReturn(paginatedTables.subList(0, pageSize));
            } else {
                when(mockResult.getNextToken()).thenReturn(null);
                when(mockResult.getTableList()).thenReturn(paginatedTables);
            }
        }
        return mockResult;
    });
}

Example 3 with GetSplitsResponse

use of com.amazonaws.athena.connector.lambda.metadata.GetSplitsResponse in project aws-athena-query-federation by awslabs.

the class ExampleMetadataHandlerTest method doGetSplits.

/**
 * The goal of this test is to test happy case for getting splits and also to exercise the continuation token
 * logic specifically.
 */
@Test
public void doGetSplits() {
    logger.info("doGetSplits: enter");
    String yearCol = "year";
    String monthCol = "month";
    String dayCol = "day";
    // This is the schema that ExampleMetadataHandler has layed out for a 'Partition' so we need to populate this
    // minimal set of info here.
    Schema schema = SchemaBuilder.newBuilder().addField(yearCol, new ArrowType.Int(16, false)).addField(monthCol, new ArrowType.Int(16, false)).addField(dayCol, new ArrowType.Int(16, false)).addField(ExampleMetadataHandler.PARTITION_LOCATION, new ArrowType.Utf8()).addField(ExampleMetadataHandler.SERDE, new ArrowType.Utf8()).build();
    List<String> partitionCols = new ArrayList<>();
    partitionCols.add(yearCol);
    partitionCols.add(monthCol);
    partitionCols.add(dayCol);
    Map<String, ValueSet> constraintsMap = new HashMap<>();
    constraintsMap.put(dayCol, SortedRangeSet.copyOf(Types.MinorType.INT.getType(), ImmutableList.of(Range.greaterThan(allocator, Types.MinorType.INT.getType(), 20)), false));
    Block partitions = allocator.createBlock(schema);
    int num_partitions = 100;
    for (int i = 0; i < num_partitions; i++) {
        BlockUtils.setValue(partitions.getFieldVector(yearCol), i, 2016 + i);
        BlockUtils.setValue(partitions.getFieldVector(monthCol), i, (i % 12) + 1);
        BlockUtils.setValue(partitions.getFieldVector(dayCol), i, (i % 28) + 1);
        BlockUtils.setValue(partitions.getFieldVector(ExampleMetadataHandler.PARTITION_LOCATION), i, String.valueOf(i));
        BlockUtils.setValue(partitions.getFieldVector(ExampleMetadataHandler.SERDE), i, "TextInputType");
    }
    partitions.setRowCount(num_partitions);
    String continuationToken = null;
    GetSplitsRequest originalReq = new GetSplitsRequest(IdentityUtil.fakeIdentity(), "queryId", "catalog_name", new TableName("schema", "table_name"), partitions, partitionCols, new Constraints(constraintsMap), continuationToken);
    int numContinuations = 0;
    do {
        GetSplitsRequest req = new GetSplitsRequest(originalReq, continuationToken);
        ObjectMapperUtil.assertSerialization(req);
        logger.info("doGetSplits: req[{}]", req);
        metadataHandler.setEncryption(numContinuations % 2 == 0);
        logger.info("doGetSplits: Toggle encryption " + (numContinuations % 2 == 0));
        MetadataResponse rawResponse = metadataHandler.doGetSplits(allocator, req);
        ObjectMapperUtil.assertSerialization(rawResponse);
        assertEquals(MetadataRequestType.GET_SPLITS, rawResponse.getRequestType());
        GetSplitsResponse response = (GetSplitsResponse) rawResponse;
        continuationToken = response.getContinuationToken();
        logger.info("doGetSplits: continuationToken[{}] - numSplits[{}] - maxSplits[{}]", new Object[] { continuationToken, response.getSplits().size(), MAX_SPLITS_PER_REQUEST });
        for (Split nextSplit : response.getSplits()) {
            if (numContinuations % 2 == 0) {
                assertNotNull(nextSplit.getEncryptionKey());
            } else {
                assertNull(nextSplit.getEncryptionKey());
            }
            assertNotNull(nextSplit.getProperty(SplitProperties.LOCATION.getId()));
            assertNotNull(nextSplit.getProperty(SplitProperties.SERDE.getId()));
            assertNotNull(nextSplit.getProperty(SplitProperties.SPLIT_PART.getId()));
        }
        assertTrue("Continuation criteria violated", (response.getSplits().size() == MAX_SPLITS_PER_REQUEST && response.getContinuationToken() != null) || response.getSplits().size() < MAX_SPLITS_PER_REQUEST);
        if (continuationToken != null) {
            numContinuations++;
        }
    } while (continuationToken != null);
    assertTrue(numContinuations > 0);
    logger.info("doGetSplits: exit");
}

Example 4 with GetSplitsResponse

use of com.amazonaws.athena.connector.lambda.metadata.GetSplitsResponse in project aws-athena-query-federation by awslabs.

the class GetSplitsResponseSerDeTest method beforeTest.

@Before
public void beforeTest() throws IOException {
    String yearCol = "year";
    String monthCol = "month";
    String dayCol = "day";
    SpillLocation spillLocation1 = S3SpillLocation.newBuilder().withBucket("athena-virtuoso-test").withPrefix("lambda-spill").withQueryId("test-query-id").withSplitId("test-split-id-1").withIsDirectory(true).build();
    EncryptionKey encryptionKey1 = new EncryptionKey("test-key-1".getBytes(), "test-nonce-1".getBytes());
    Split split1 = Split.newBuilder(spillLocation1, encryptionKey1).add(yearCol, "2017").add(monthCol, "11").add(dayCol, "1").build();
    SpillLocation spillLocation2 = S3SpillLocation.newBuilder().withBucket("athena-virtuoso-test").withPrefix("lambda-spill").withQueryId("test-query-id").withSplitId("test-split-id-2").withIsDirectory(true).build();
    EncryptionKey encryptionKey2 = new EncryptionKey("test-key-2".getBytes(), "test-nonce-2".getBytes());
    Split split2 = Split.newBuilder(spillLocation2, encryptionKey2).add(yearCol, "2017").add(monthCol, "11").add(dayCol, "2").build();
    expected = new GetSplitsResponse("test-catalog", ImmutableSet.of(split1, split2), "test-continuation-token");
    String expectedSerDeFile = utils.getResourceOrFail("serde/v2", "GetSplitsResponse.json");
    expectedSerDeText = utils.readAllAsString(expectedSerDeFile).trim();
}

Example 5 with GetSplitsResponse

use of com.amazonaws.athena.connector.lambda.metadata.GetSplitsResponse in project aws-athena-query-federation by awslabs.

the class GetSplitsResponseSerDeTest method deserialize.

@Test
public void deserialize() throws IOException {
    logger.info("deserialize: enter");
    InputStream input = new ByteArrayInputStream(expectedSerDeText.getBytes());
    GetSplitsResponse actual = (GetSplitsResponse) mapper.readValue(input, FederationResponse.class);
    logger.info("deserialize: deserialized[{}]", actual);
    assertEquals(expected, actual);
    logger.info("deserialize: exit");
}