Search in sources :

Example 16 with ConnectorSplit

use of com.facebook.presto.spi.ConnectorSplit in project presto by prestodb.

the class AbstractTestHiveClientS3 method testGetRecordsS3.

@Test
public void testGetRecordsS3() throws Exception {
    try (Transaction transaction = newTransaction()) {
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorSession session = newSession();
        ConnectorTableHandle table = getTableHandle(metadata, tableS3);
        List<ColumnHandle> columnHandles = ImmutableList.copyOf(metadata.getColumnHandles(session, table).values());
        Map<String, Integer> columnIndex = indexColumns(columnHandles);
        List<ConnectorTableLayoutResult> tableLayoutResults = metadata.getTableLayouts(session, table, new Constraint<>(TupleDomain.all(), bindings -> true), Optional.empty());
        HiveTableLayoutHandle layoutHandle = (HiveTableLayoutHandle) getOnlyElement(tableLayoutResults).getTableLayout().getHandle();
        assertEquals(layoutHandle.getPartitions().get().size(), 1);
        ConnectorSplitSource splitSource = splitManager.getSplits(transaction.getTransactionHandle(), session, layoutHandle);
        long sum = 0;
        for (ConnectorSplit split : getAllSplits(splitSource)) {
            try (ConnectorPageSource pageSource = pageSourceProvider.createPageSource(transaction.getTransactionHandle(), session, split, columnHandles)) {
                MaterializedResult result = materializeSourceDataStream(session, pageSource, getTypes(columnHandles));
                for (MaterializedRow row : result) {
                    sum += (Long) row.getField(columnIndex.get("t_bigint"));
                }
            }
        }
        assertEquals(sum, 78300);
    }
}
Also used : HiveTestUtils.getDefaultHiveDataStreamFactories(com.facebook.presto.hive.HiveTestUtils.getDefaultHiveDataStreamFactories) ConnectorSplitSource(com.facebook.presto.spi.ConnectorSplitSource) HiveTestUtils.getDefaultHiveFileWriterFactories(com.facebook.presto.hive.HiveTestUtils.getDefaultHiveFileWriterFactories) FileSystem(org.apache.hadoop.fs.FileSystem) TypeRegistry(com.facebook.presto.type.TypeRegistry) Test(org.testng.annotations.Test) AbstractTestHiveClient.getAllSplits(com.facebook.presto.hive.AbstractTestHiveClient.getAllSplits) BIGINT(com.facebook.presto.spi.type.BigintType.BIGINT) MaterializedResult.materializeSourceDataStream(com.facebook.presto.testing.MaterializedResult.materializeSourceDataStream) Preconditions.checkArgument(com.google.common.base.Preconditions.checkArgument) SchemaTableName(com.facebook.presto.spi.SchemaTableName) ExtendedHiveMetastore(com.facebook.presto.hive.metastore.ExtendedHiveMetastore) BoundedExecutor(io.airlift.concurrent.BoundedExecutor) Map(java.util.Map) ConnectorPageSink(com.facebook.presto.spi.ConnectorPageSink) Path(org.apache.hadoop.fs.Path) ConnectorPageSourceProvider(com.facebook.presto.spi.connector.ConnectorPageSourceProvider) ENGLISH(java.util.Locale.ENGLISH) Assert.assertFalse(org.testng.Assert.assertFalse) AbstractTestHiveClient.filterNonHiddenColumnHandles(com.facebook.presto.hive.AbstractTestHiveClient.filterNonHiddenColumnHandles) ConnectorSplitManager(com.facebook.presto.spi.connector.ConnectorSplitManager) ImmutableMap(com.google.common.collect.ImmutableMap) PrincipalPrivileges(com.facebook.presto.hive.metastore.PrincipalPrivileges) BeforeClass(org.testng.annotations.BeforeClass) Collection(java.util.Collection) UUID(java.util.UUID) String.format(java.lang.String.format) ConnectorSession(com.facebook.presto.spi.ConnectorSession) TupleDomain(com.facebook.presto.spi.predicate.TupleDomain) List(java.util.List) ColumnMetadata(com.facebook.presto.spi.ColumnMetadata) BridgingHiveMetastore(com.facebook.presto.hive.metastore.BridgingHiveMetastore) TYPE_MANAGER(com.facebook.presto.hive.HiveTestUtils.TYPE_MANAGER) Optional(java.util.Optional) NoHdfsAuthentication(com.facebook.presto.hive.authentication.NoHdfsAuthentication) JsonCodec(io.airlift.json.JsonCodec) ConnectorMetadata(com.facebook.presto.spi.connector.ConnectorMetadata) Table(com.facebook.presto.hive.metastore.Table) Slice(io.airlift.slice.Slice) HiveTransaction(com.facebook.presto.hive.AbstractTestHiveClient.HiveTransaction) Database(com.facebook.presto.hive.metastore.Database) HiveTestUtils.getTypes(com.facebook.presto.hive.HiveTestUtils.getTypes) MoreExecutors.newDirectExecutorService(com.google.common.util.concurrent.MoreExecutors.newDirectExecutorService) Assert.assertEquals(org.testng.Assert.assertEquals) ConnectorTableHandle(com.facebook.presto.spi.ConnectorTableHandle) ConnectorOutputTableHandle(com.facebook.presto.spi.ConnectorOutputTableHandle) ImmutableList(com.google.common.collect.ImmutableList) Threads.daemonThreadsNamed(io.airlift.concurrent.Threads.daemonThreadsNamed) ConnectorPageSinkProvider(com.facebook.presto.spi.connector.ConnectorPageSinkProvider) ImmutableMultimap(com.google.common.collect.ImmutableMultimap) AbstractTestHiveClient.createTableProperties(com.facebook.presto.hive.AbstractTestHiveClient.createTableProperties) ExecutorService(java.util.concurrent.ExecutorService) ConnectorTableMetadata(com.facebook.presto.spi.ConnectorTableMetadata) AfterClass(org.testng.annotations.AfterClass) CachingHiveMetastore(com.facebook.presto.hive.metastore.CachingHiveMetastore) Transaction(com.facebook.presto.hive.AbstractTestHiveClient.Transaction) Throwables(com.google.common.base.Throwables) Constraint(com.facebook.presto.spi.Constraint) Iterables.getOnlyElement(com.google.common.collect.Iterables.getOnlyElement) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) MoreFutures.getFutureValue(io.airlift.concurrent.MoreFutures.getFutureValue) HostAndPort(com.google.common.net.HostAndPort) ConnectorSplit(com.facebook.presto.spi.ConnectorSplit) ConnectorTableLayoutResult(com.facebook.presto.spi.ConnectorTableLayoutResult) MaterializedResult(com.facebook.presto.testing.MaterializedResult) Assertions.assertEqualsIgnoreOrder(io.airlift.testing.Assertions.assertEqualsIgnoreOrder) ConnectorPageSource(com.facebook.presto.spi.ConnectorPageSource) TableNotFoundException(com.facebook.presto.spi.TableNotFoundException) ColumnHandle(com.facebook.presto.spi.ColumnHandle) Executors.newCachedThreadPool(java.util.concurrent.Executors.newCachedThreadPool) MaterializedRow(com.facebook.presto.testing.MaterializedRow) AbstractTestHiveClient.filterNonHiddenColumnMetadata(com.facebook.presto.hive.AbstractTestHiveClient.filterNonHiddenColumnMetadata) Assert.assertTrue(org.testng.Assert.assertTrue) HadoopFileStatus.isDirectory(com.facebook.presto.hadoop.HadoopFileStatus.isDirectory) GroupByHashPageIndexerFactory(com.facebook.presto.GroupByHashPageIndexerFactory) ThriftHiveMetastore(com.facebook.presto.hive.metastore.ThriftHiveMetastore) JoinCompiler(com.facebook.presto.sql.gen.JoinCompiler) HiveTestUtils.getDefaultHiveRecordCursorProvider(com.facebook.presto.hive.HiveTestUtils.getDefaultHiveRecordCursorProvider) ColumnHandle(com.facebook.presto.spi.ColumnHandle) ConnectorSplitSource(com.facebook.presto.spi.ConnectorSplitSource) ConnectorPageSource(com.facebook.presto.spi.ConnectorPageSource) ConnectorTableHandle(com.facebook.presto.spi.ConnectorTableHandle) HiveTransaction(com.facebook.presto.hive.AbstractTestHiveClient.HiveTransaction) Transaction(com.facebook.presto.hive.AbstractTestHiveClient.Transaction) ConnectorTableLayoutResult(com.facebook.presto.spi.ConnectorTableLayoutResult) ConnectorSession(com.facebook.presto.spi.ConnectorSession) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) ConnectorMetadata(com.facebook.presto.spi.connector.ConnectorMetadata) MaterializedResult(com.facebook.presto.testing.MaterializedResult) ConnectorSplit(com.facebook.presto.spi.ConnectorSplit) MaterializedRow(com.facebook.presto.testing.MaterializedRow) Test(org.testng.annotations.Test)

Example 17 with ConnectorSplit

use of com.facebook.presto.spi.ConnectorSplit in project presto by prestodb.

the class NodePartitioningManager method getNodePartitioningMap.

public NodePartitionMap getNodePartitioningMap(Session session, PartitioningHandle partitioningHandle) {
    requireNonNull(session, "session is null");
    requireNonNull(partitioningHandle, "partitioningHandle is null");
    if (partitioningHandle.getConnectorHandle() instanceof SystemPartitioningHandle) {
        return ((SystemPartitioningHandle) partitioningHandle.getConnectorHandle()).getNodePartitionMap(session, nodeScheduler);
    }
    ConnectorNodePartitioningProvider partitioningProvider = partitioningProviders.get(partitioningHandle.getConnectorId().get());
    checkArgument(partitioningProvider != null, "No partitioning provider for connector %s", partitioningHandle.getConnectorId().get());
    Map<Integer, Node> bucketToNode = partitioningProvider.getBucketToNode(partitioningHandle.getTransactionHandle().orElse(null), session.toConnectorSession(), partitioningHandle.getConnectorHandle());
    checkArgument(bucketToNode != null, "No partition map %s", partitioningHandle);
    checkArgument(!bucketToNode.isEmpty(), "Partition map %s is empty", partitioningHandle);
    int bucketCount = bucketToNode.keySet().stream().mapToInt(Integer::intValue).max().getAsInt() + 1;
    // safety check for crazy partitioning
    checkArgument(bucketCount < 1_000_000, "Too many buckets in partitioning: %s", bucketCount);
    int[] bucketToPartition = new int[bucketCount];
    BiMap<Node, Integer> nodeToPartition = HashBiMap.create();
    int nextPartitionId = 0;
    for (Entry<Integer, Node> entry : bucketToNode.entrySet()) {
        Integer partitionId = nodeToPartition.get(entry.getValue());
        if (partitionId == null) {
            partitionId = nextPartitionId++;
            nodeToPartition.put(entry.getValue(), partitionId);
        }
        bucketToPartition[entry.getKey()] = partitionId;
    }
    ToIntFunction<ConnectorSplit> splitBucketFunction = partitioningProvider.getSplitBucketFunction(partitioningHandle.getTransactionHandle().orElse(null), session.toConnectorSession(), partitioningHandle.getConnectorHandle());
    checkArgument(splitBucketFunction != null, "No partitioning %s", partitioningHandle);
    return new NodePartitionMap(nodeToPartition.inverse(), bucketToPartition, split -> splitBucketFunction.applyAsInt(split.getConnectorSplit()));
}
Also used : ConnectorNodePartitioningProvider(com.facebook.presto.spi.connector.ConnectorNodePartitioningProvider) Node(com.facebook.presto.spi.Node) ConnectorSplit(com.facebook.presto.spi.ConnectorSplit)

Example 18 with ConnectorSplit

use of com.facebook.presto.spi.ConnectorSplit in project presto by prestodb.

the class MemorySplitManager method getSplits.

@Override
public ConnectorSplitSource getSplits(ConnectorTransactionHandle transactionHandle, ConnectorSession session, ConnectorTableLayoutHandle layoutHandle) {
    MemoryTableLayoutHandle layout = (MemoryTableLayoutHandle) layoutHandle;
    List<HostAddress> hosts = layout.getTable().getHosts();
    ImmutableList.Builder<ConnectorSplit> splits = ImmutableList.builder();
    for (HostAddress host : hosts) {
        for (int i = 0; i < splitsPerNode; i++) {
            splits.add(new MemorySplit(layout.getTable(), i, splitsPerNode, ImmutableList.of(host)));
        }
    }
    return new FixedSplitSource(splits.build());
}
Also used : ImmutableList(com.google.common.collect.ImmutableList) FixedSplitSource(com.facebook.presto.spi.FixedSplitSource) HostAddress(com.facebook.presto.spi.HostAddress) ConnectorSplit(com.facebook.presto.spi.ConnectorSplit)

Example 19 with ConnectorSplit

use of com.facebook.presto.spi.ConnectorSplit in project presto by prestodb.

the class SystemSplitManager method getSplits.

@Override
public ConnectorSplitSource getSplits(ConnectorTransactionHandle transactionHandle, ConnectorSession session, ConnectorTableLayoutHandle layout) {
    SystemTableLayoutHandle layoutHandle = (SystemTableLayoutHandle) layout;
    SystemTableHandle tableHandle = layoutHandle.getTable();
    TupleDomain<ColumnHandle> constraint = layoutHandle.getConstraint();
    SystemTable systemTable = tables.get(tableHandle.getSchemaTableName());
    Distribution tableDistributionMode = systemTable.getDistribution();
    if (tableDistributionMode == SINGLE_COORDINATOR) {
        HostAddress address = nodeManager.getCurrentNode().getHostAndPort();
        ConnectorSplit split = new SystemSplit(tableHandle.getConnectorId(), tableHandle, address, constraint);
        return new FixedSplitSource(ImmutableList.of(split));
    }
    ImmutableList.Builder<ConnectorSplit> splits = ImmutableList.builder();
    ImmutableSet.Builder<Node> nodes = ImmutableSet.builder();
    if (tableDistributionMode == ALL_COORDINATORS) {
        nodes.addAll(nodeManager.getCoordinators());
    } else if (tableDistributionMode == ALL_NODES) {
        nodes.addAll(nodeManager.getNodes(ACTIVE));
    }
    Set<Node> nodeSet = nodes.build();
    for (Node node : nodeSet) {
        splits.add(new SystemSplit(tableHandle.getConnectorId(), tableHandle, node.getHostAndPort(), constraint));
    }
    return new FixedSplitSource(splits.build());
}
Also used : ColumnHandle(com.facebook.presto.spi.ColumnHandle) ImmutableList(com.google.common.collect.ImmutableList) Node(com.facebook.presto.spi.Node) HostAddress(com.facebook.presto.spi.HostAddress) ImmutableSet(com.google.common.collect.ImmutableSet) FixedSplitSource(com.facebook.presto.spi.FixedSplitSource) Distribution(com.facebook.presto.spi.SystemTable.Distribution) SystemTable(com.facebook.presto.spi.SystemTable) ConnectorSplit(com.facebook.presto.spi.ConnectorSplit)

Example 20 with ConnectorSplit

use of com.facebook.presto.spi.ConnectorSplit in project presto by prestodb.

the class KafkaSplitManager method getSplits.

@Override
public ConnectorSplitSource getSplits(ConnectorTransactionHandle transaction, ConnectorSession session, ConnectorTableLayoutHandle layout) {
    KafkaTableHandle kafkaTableHandle = convertLayout(layout).getTable();
    SimpleConsumer simpleConsumer = consumerManager.getConsumer(selectRandom(nodes));
    TopicMetadataRequest topicMetadataRequest = new TopicMetadataRequest(ImmutableList.of(kafkaTableHandle.getTopicName()));
    TopicMetadataResponse topicMetadataResponse = simpleConsumer.send(topicMetadataRequest);
    ImmutableList.Builder<ConnectorSplit> splits = ImmutableList.builder();
    for (TopicMetadata metadata : topicMetadataResponse.topicsMetadata()) {
        for (PartitionMetadata part : metadata.partitionsMetadata()) {
            log.debug("Adding Partition %s/%s", metadata.topic(), part.partitionId());
            Broker leader = part.leader();
            if (leader == null) {
                // Leader election going on...
                log.warn("No leader for partition %s/%s found!", metadata.topic(), part.partitionId());
                continue;
            }
            HostAddress partitionLeader = HostAddress.fromParts(leader.host(), leader.port());
            SimpleConsumer leaderConsumer = consumerManager.getConsumer(partitionLeader);
            // Kafka contains a reverse list of "end - start" pairs for the splits
            long[] offsets = findAllOffsets(leaderConsumer, metadata.topic(), part.partitionId());
            for (int i = offsets.length - 1; i > 0; i--) {
                KafkaSplit split = new KafkaSplit(connectorId, metadata.topic(), kafkaTableHandle.getKeyDataFormat(), kafkaTableHandle.getMessageDataFormat(), part.partitionId(), offsets[i], offsets[i - 1], partitionLeader);
                splits.add(split);
            }
        }
    }
    return new FixedSplitSource(splits.build());
}
Also used : Broker(kafka.cluster.Broker) ImmutableList(com.google.common.collect.ImmutableList) TopicMetadataRequest(kafka.javaapi.TopicMetadataRequest) TopicMetadataResponse(kafka.javaapi.TopicMetadataResponse) HostAddress(com.facebook.presto.spi.HostAddress) TopicMetadata(kafka.javaapi.TopicMetadata) FixedSplitSource(com.facebook.presto.spi.FixedSplitSource) PartitionMetadata(kafka.javaapi.PartitionMetadata) ConnectorSplit(com.facebook.presto.spi.ConnectorSplit) SimpleConsumer(kafka.javaapi.consumer.SimpleConsumer)

Aggregations

ConnectorSplit (com.facebook.presto.spi.ConnectorSplit)31 ColumnHandle (com.facebook.presto.spi.ColumnHandle)15 ImmutableList (com.google.common.collect.ImmutableList)15 Test (org.testng.annotations.Test)13 ConnectorSession (com.facebook.presto.spi.ConnectorSession)11 ConnectorSplitSource (com.facebook.presto.spi.ConnectorSplitSource)11 FixedSplitSource (com.facebook.presto.spi.FixedSplitSource)10 List (java.util.List)10 ConnectorPageSource (com.facebook.presto.spi.ConnectorPageSource)9 ConnectorTableHandle (com.facebook.presto.spi.ConnectorTableHandle)9 ConnectorTableLayoutHandle (com.facebook.presto.spi.ConnectorTableLayoutHandle)9 Constraint (com.facebook.presto.spi.Constraint)9 ConnectorMetadata (com.facebook.presto.spi.connector.ConnectorMetadata)8 TupleDomain (com.facebook.presto.spi.predicate.TupleDomain)8 MaterializedResult (com.facebook.presto.testing.MaterializedResult)8 MaterializedRow (com.facebook.presto.testing.MaterializedRow)8 TestingConnectorSession (com.facebook.presto.testing.TestingConnectorSession)8 ConnectorTableLayoutResult (com.facebook.presto.spi.ConnectorTableLayoutResult)7 ConnectorTableMetadata (com.facebook.presto.spi.ConnectorTableMetadata)7 ConnectorSplitManager (com.facebook.presto.spi.connector.ConnectorSplitManager)7