Examples with LoadFileType org.apache.hadoop.hive.ql.plan.LoadTableDesc.LoadFileType used on opensource projects

Example 6 with LoadFileType

use of org.apache.hadoop.hive.ql.plan.LoadTableDesc.LoadFileType in project hive by apache.

the class Hive method loadDynamicPartitions.

/**
 * Given a source directory name of the load path, load all dynamically generated partitions
 * into the specified table and return a list of strings that represent the dynamic partition
 * paths.
 * @param tbd table descriptor
 * @param numLB number of buckets
 * @param isAcid true if this is an ACID operation
 * @param writeId writeId, can be 0 unless isAcid == true
 * @param stmtId statementId
 * @param resetStatistics if true, reset statistics. Do not reset statistics otherwise.
 * @param operation ACID operation type
 * @param partitionDetailsMap full dynamic partition specification
 * @return partition map details (PartitionSpec and Partition)
 * @throws HiveException
 */
public Map<Map<String, String>, Partition> loadDynamicPartitions(final LoadTableDesc tbd, final int numLB, final boolean isAcid, final long writeId, final int stmtId, final boolean resetStatistics, final AcidUtils.Operation operation, Map<Path, PartitionDetails> partitionDetailsMap) throws HiveException {
    PerfLogger perfLogger = SessionState.getPerfLogger();
    perfLogger.perfLogBegin("MoveTask", PerfLogger.LOAD_DYNAMIC_PARTITIONS);
    final Path loadPath = tbd.getSourcePath();
    final Table tbl = getTable(tbd.getTable().getTableName());
    final Map<String, String> partSpec = tbd.getPartitionSpec();
    final AtomicInteger partitionsLoaded = new AtomicInteger(0);
    final boolean inPlaceEligible = conf.getLong("fs.trash.interval", 0) <= 0 && InPlaceUpdate.canRenderInPlace(conf) && !SessionState.getConsole().getIsSilent();
    final PrintStream ps = (inPlaceEligible) ? SessionState.getConsole().getInfoStream() : null;
    final SessionState parentSession = SessionState.get();
    List<Callable<Partition>> tasks = Lists.newLinkedList();
    boolean fetchPartitionInfo = true;
    final boolean scanPartitionsByName = HiveConf.getBoolVar(conf, HIVE_LOAD_DYNAMIC_PARTITIONS_SCAN_SPECIFIC_PARTITIONS);
    // for every dynamic partition
    if (scanPartitionsByName && !tbd.isDirectInsert() && !AcidUtils.isTransactionalTable(tbl)) {
        // Fetch only relevant partitions from HMS for checking old partitions
        List<String> partitionNames = new LinkedList<>();
        for (PartitionDetails details : partitionDetailsMap.values()) {
            if (details.fullSpec != null && !details.fullSpec.isEmpty()) {
                partitionNames.add(Warehouse.makeDynamicPartNameNoTrailingSeperator(details.fullSpec));
            }
        }
        List<Partition> partitions = Hive.get().getPartitionsByNames(tbl, partitionNames);
        for (Partition partition : partitions) {
            LOG.debug("HMS partition spec: {}", partition.getSpec());
            partitionDetailsMap.entrySet().parallelStream().filter(entry -> entry.getValue().fullSpec.equals(partition.getSpec())).findAny().ifPresent(entry -> {
                entry.getValue().partition = partition;
                entry.getValue().hasOldPartition = true;
            });
        }
        // no need to fetch partition again in tasks since we have already fetched partitions
        // info in getPartitionsByNames()
        fetchPartitionInfo = false;
    }
    boolean isTxnTable = AcidUtils.isTransactionalTable(tbl);
    AcidUtils.TableSnapshot tableSnapshot = isTxnTable ? getTableSnapshot(tbl, writeId) : null;
    for (Entry<Path, PartitionDetails> entry : partitionDetailsMap.entrySet()) {
        boolean getPartitionFromHms = fetchPartitionInfo;
        tasks.add(() -> {
            PartitionDetails partitionDetails = entry.getValue();
            Map<String, String> fullPartSpec = partitionDetails.fullSpec;
            try {
                SessionState.setCurrentSessionState(parentSession);
                if (getPartitionFromHms) {
                    // didn't fetch partition info from HMS. Getting from HMS now.
                    Partition existing = getPartition(tbl, fullPartSpec, false);
                    if (existing != null) {
                        partitionDetails.partition = existing;
                        partitionDetails.hasOldPartition = true;
                    }
                }
                LOG.info("New loading path = " + entry.getKey() + " withPartSpec " + fullPartSpec);
                Partition oldPartition = partitionDetails.partition;
                List<FileStatus> newFiles = null;
                if (partitionDetails.newFiles != null) {
                    // If we already know the files from the direct insert manifest, use them
                    newFiles = partitionDetails.newFiles;
                } else if (conf.getBoolVar(ConfVars.FIRE_EVENTS_FOR_DML) && !tbl.isTemporary() && oldPartition == null) {
                    // Otherwise only collect them, if we are going to fire write notifications
                    newFiles = Collections.synchronizedList(new ArrayList<>());
                }
                // load the partition
                Partition partition = loadPartitionInternal(entry.getKey(), tbl, fullPartSpec, oldPartition, tbd.getLoadFileType(), true, false, numLB > 0, false, isAcid, resetStatistics, writeId, stmtId, tbd.isInsertOverwrite(), isTxnTable, newFiles, tbd.isDirectInsert());
                // metastore
                if (tableSnapshot != null) {
                    partition.getTPartition().setWriteId(tableSnapshot.getWriteId());
                }
                partitionDetails.tableSnapshot = tableSnapshot;
                if (oldPartition == null) {
                    partitionDetails.newFiles = newFiles;
                    partitionDetails.partition = partition;
                }
                if (inPlaceEligible) {
                    synchronized (ps) {
                        InPlaceUpdate.rePositionCursor(ps);
                        partitionsLoaded.incrementAndGet();
                        InPlaceUpdate.reprintLine(ps, "Loaded : " + partitionsLoaded.get() + "/" + partitionDetailsMap.size() + " partitions.");
                    }
                }
                return partition;
            } catch (Exception e) {
                LOG.error("Exception when loading partition with parameters " + " partPath=" + entry.getKey() + ", " + " table=" + tbl.getTableName() + ", " + " partSpec=" + fullPartSpec + ", " + " loadFileType=" + tbd.getLoadFileType().toString() + ", " + " listBucketingLevel=" + numLB + ", " + " isAcid=" + isAcid + ", " + " resetStatistics=" + resetStatistics, e);
                throw e;
            } finally {
                // get(conf).getMSC can be called in this task, Close the HMS connection right after use, do not wait for finalizer to close it.
                closeCurrent();
            }
        });
    }
    int poolSize = conf.getInt(ConfVars.HIVE_LOAD_DYNAMIC_PARTITIONS_THREAD_COUNT.varname, 1);
    ExecutorService executor = Executors.newFixedThreadPool(poolSize, new ThreadFactoryBuilder().setDaemon(true).setNameFormat("load-dynamic-partitionsToAdd-%d").build());
    List<Future<Partition>> futures = Lists.newLinkedList();
    Map<Map<String, String>, Partition> result = Maps.newLinkedHashMap();
    try {
        futures = executor.invokeAll(tasks);
        LOG.info("Number of partitionsToAdd to be added is " + futures.size());
        for (Future<Partition> future : futures) {
            Partition partition = future.get();
            result.put(partition.getSpec(), partition);
        }
        // add new partitions in batch
        addPartitionsToMetastore(partitionDetailsMap.entrySet().stream().filter(entry -> !entry.getValue().hasOldPartition).map(entry -> entry.getValue().partition).collect(Collectors.toList()), resetStatistics, tbl, partitionDetailsMap.entrySet().stream().filter(entry -> !entry.getValue().hasOldPartition).map(entry -> entry.getValue().tableSnapshot).collect(Collectors.toList()));
        // For acid table, add the acid_write event with file list at the time of load itself. But
        // it should be done after partition is created.
        List<WriteNotificationLogRequest> requestList = new ArrayList<>();
        int maxBatchSize = conf.getIntVar(HIVE_WRITE_NOTIFICATION_MAX_BATCH_SIZE);
        for (Entry<Path, PartitionDetails> entry : partitionDetailsMap.entrySet()) {
            PartitionDetails partitionDetails = entry.getValue();
            if (isTxnTable && partitionDetails.newFiles != null) {
                addWriteNotificationLog(tbl, partitionDetails.fullSpec, partitionDetails.newFiles, writeId, requestList);
                if (requestList != null && requestList.size() >= maxBatchSize) {
                    // If the first call returns that the HMS does not supports batching, avoid batching
                    // for later requests.
                    boolean batchSupported = addWriteNotificationLogInBatch(tbl, requestList);
                    if (batchSupported) {
                        requestList.clear();
                    } else {
                        requestList = null;
                    }
                }
            }
        }
        if (requestList != null && requestList.size() > 0) {
            addWriteNotificationLogInBatch(tbl, requestList);
        }
        setStatsPropAndAlterPartitions(resetStatistics, tbl, partitionDetailsMap.entrySet().stream().filter(entry -> entry.getValue().hasOldPartition).map(entry -> entry.getValue().partition).collect(Collectors.toList()), tableSnapshot);
    } catch (InterruptedException | ExecutionException e) {
        throw new HiveException("Exception when loading " + partitionDetailsMap.size() + " partitions" + " in table " + tbl.getTableName() + " with loadPath=" + loadPath, e);
    } catch (TException e) {
        LOG.error("Failed loadDynamicPartitions", e);
        throw new HiveException(e);
    } catch (Exception e) {
        StringBuffer logMsg = new StringBuffer();
        logMsg.append("Exception when loading partitionsToAdd with parameters ");
        logMsg.append("partPaths=");
        partitionDetailsMap.keySet().forEach(path -> logMsg.append(path + ", "));
        logMsg.append("table=" + tbl.getTableName() + ", ").append("partSpec=" + partSpec + ", ").append("loadFileType=" + tbd.getLoadFileType().toString() + ", ").append("listBucketingLevel=" + numLB + ", ").append("isAcid=" + isAcid + ", ").append("resetStatistics=" + resetStatistics);
        LOG.error(logMsg.toString(), e);
        throw e;
    } finally {
        LOG.debug("Cancelling " + futures.size() + " dynamic loading tasks");
        executor.shutdownNow();
    }
    if (HiveConf.getBoolVar(conf, ConfVars.HIVE_IN_TEST) && HiveConf.getBoolVar(conf, ConfVars.HIVETESTMODEFAILLOADDYNAMICPARTITION)) {
        throw new HiveException(HiveConf.ConfVars.HIVETESTMODEFAILLOADDYNAMICPARTITION.name() + "=true");
    }
    try {
        if (isTxnTable) {
            List<String> partNames = result.values().stream().map(Partition::getName).collect(Collectors.toList());
            getMSC().addDynamicPartitions(parentSession.getTxnMgr().getCurrentTxnId(), writeId, tbl.getDbName(), tbl.getTableName(), partNames, AcidUtils.toDataOperationType(operation));
        }
        LOG.info("Loaded " + result.size() + "partitionsToAdd");
        perfLogger.perfLogEnd("MoveTask", PerfLogger.LOAD_DYNAMIC_PARTITIONS);
        return result;
    } catch (TException te) {
        LOG.error("Failed loadDynamicPartitions", te);
        throw new HiveException("Exception updating metastore for acid table " + tbd.getTable().getTableName() + " with partitions " + result.values(), te);
    }
}