Examples with PrestoException com.facebook.presto.spi.PrestoException used on opensource projects

Example 26 with PrestoException

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

the class HiveMetadata method beginInsert.

@Override
public HiveInsertTableHandle beginInsert(ConnectorSession session, ConnectorTableHandle tableHandle) {
    verifyJvmTimeZone();
    SchemaTableName tableName = schemaTableName(tableHandle);
    Optional<Table> table = metastore.getTable(tableName.getSchemaName(), tableName.getTableName());
    if (!table.isPresent()) {
        throw new TableNotFoundException(tableName);
    }
    checkTableIsWritable(table.get());
    for (Column column : table.get().getDataColumns()) {
        if (!isWritableType(column.getType())) {
            throw new PrestoException(NOT_SUPPORTED, format("Inserting into Hive table %s.%s with column type %s not supported", table.get().getDatabaseName(), table.get().getTableName(), column.getType()));
        }
    }
    List<HiveColumnHandle> handles = hiveColumnHandles(connectorId, table.get()).stream().filter(columnHandle -> !columnHandle.isHidden()).collect(toList());
    HiveStorageFormat tableStorageFormat = extractHiveStorageFormat(table.get());
    LocationHandle locationHandle = locationService.forExistingTable(metastore, session.getUser(), session.getQueryId(), table.get());
    HiveInsertTableHandle result = new HiveInsertTableHandle(connectorId, tableName.getSchemaName(), tableName.getTableName(), handles, session.getQueryId(), metastore.generatePageSinkMetadata(tableName), locationHandle, table.get().getStorage().getBucketProperty(), tableStorageFormat, respectTableFormat ? tableStorageFormat : defaultStorageFormat);
    Optional<Path> writePathRoot = locationService.writePathRoot(locationHandle);
    Path targetPathRoot = locationService.targetPathRoot(locationHandle);
    if (writePathRoot.isPresent()) {
        WriteMode mode = writePathRoot.get().equals(targetPathRoot) ? DIRECT_TO_TARGET_NEW_DIRECTORY : STAGE_AND_MOVE_TO_TARGET_DIRECTORY;
        metastore.declareIntentionToWrite(session, mode, writePathRoot.get(), result.getFilePrefix(), tableName);
    } else {
        metastore.declareIntentionToWrite(session, DIRECT_TO_TARGET_EXISTING_DIRECTORY, targetPathRoot, result.getFilePrefix(), tableName);
    }
    return result;
}

Example 27 with PrestoException

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

the class HivePartitionManager method getPartitions.

public HivePartitionResult getPartitions(SemiTransactionalHiveMetastore metastore, ConnectorTableHandle tableHandle, Constraint<ColumnHandle> constraint) {
    HiveTableHandle hiveTableHandle = (HiveTableHandle) tableHandle;
    TupleDomain<ColumnHandle> effectivePredicate = constraint.getSummary();
    SchemaTableName tableName = hiveTableHandle.getSchemaTableName();
    Table table = getTable(metastore, tableName);
    Optional<HiveBucketHandle> hiveBucketHandle = getHiveBucketHandle(connectorId, table);
    List<HiveColumnHandle> partitionColumns = getPartitionKeyColumnHandles(connectorId, table);
    List<HiveBucket> buckets = getHiveBucketNumbers(table, effectivePredicate);
    TupleDomain<HiveColumnHandle> compactEffectivePredicate = toCompactTupleDomain(effectivePredicate, domainCompactionThreshold);
    if (effectivePredicate.isNone()) {
        return new HivePartitionResult(partitionColumns, ImmutableList.of(), TupleDomain.none(), TupleDomain.none(), hiveBucketHandle);
    }
    if (partitionColumns.isEmpty()) {
        return new HivePartitionResult(partitionColumns, ImmutableList.of(new HivePartition(tableName, compactEffectivePredicate, buckets)), effectivePredicate, TupleDomain.none(), hiveBucketHandle);
    }
    List<Type> partitionTypes = partitionColumns.stream().map(column -> typeManager.getType(column.getTypeSignature())).collect(toList());
    List<String> partitionNames = getFilteredPartitionNames(metastore, tableName, partitionColumns, effectivePredicate);
    // do a final pass to filter based on fields that could not be used to filter the partitions
    int partitionCount = 0;
    ImmutableList.Builder<HivePartition> partitions = ImmutableList.builder();
    for (String partitionName : partitionNames) {
        Optional<Map<ColumnHandle, NullableValue>> values = parseValuesAndFilterPartition(partitionName, partitionColumns, partitionTypes, constraint);
        if (values.isPresent()) {
            if (partitionCount == maxPartitions) {
                throw new PrestoException(HIVE_EXCEEDED_PARTITION_LIMIT, format("Query over table '%s' can potentially read more than %s partitions", hiveTableHandle.getSchemaTableName(), maxPartitions));
            }
            partitionCount++;
            partitions.add(new HivePartition(tableName, compactEffectivePredicate, partitionName, values.get(), buckets));
        }
    }
    // All partition key domains will be fully evaluated, so we don't need to include those
    TupleDomain<ColumnHandle> remainingTupleDomain = TupleDomain.withColumnDomains(Maps.filterKeys(effectivePredicate.getDomains().get(), not(Predicates.in(partitionColumns))));
    TupleDomain<ColumnHandle> enforcedTupleDomain = TupleDomain.withColumnDomains(Maps.filterKeys(effectivePredicate.getDomains().get(), Predicates.in(partitionColumns)));
    return new HivePartitionResult(partitionColumns, partitions.build(), remainingTupleDomain, enforcedTupleDomain, hiveBucketHandle);
}

Example 28 with PrestoException

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

the class HivePartitionManager method getFilteredPartitionNames.

private List<String> getFilteredPartitionNames(SemiTransactionalHiveMetastore metastore, SchemaTableName tableName, List<HiveColumnHandle> partitionKeys, TupleDomain<ColumnHandle> effectivePredicate) {
    checkArgument(effectivePredicate.getDomains().isPresent());
    List<String> filter = new ArrayList<>();
    for (HiveColumnHandle partitionKey : partitionKeys) {
        Domain domain = effectivePredicate.getDomains().get().get(partitionKey);
        if (domain != null && domain.isNullableSingleValue()) {
            Object value = domain.getNullableSingleValue();
            if (value == null) {
                filter.add(HivePartitionKey.HIVE_DEFAULT_DYNAMIC_PARTITION);
            } else if (value instanceof Slice) {
                filter.add(((Slice) value).toStringUtf8());
            } else if ((value instanceof Boolean) || (value instanceof Double) || (value instanceof Long)) {
                if (assumeCanonicalPartitionKeys) {
                    filter.add(value.toString());
                } else {
                    // Hive treats '0', 'false', and 'False' the same. However, the metastore differentiates between these.
                    filter.add(PARTITION_VALUE_WILDCARD);
                }
            } else {
                throw new PrestoException(NOT_SUPPORTED, "Only Boolean, Double and Long partition keys are supported");
            }
        } else {
            filter.add(PARTITION_VALUE_WILDCARD);
        }
    }
    // fetch the partition names
    return metastore.getPartitionNamesByParts(tableName.getSchemaName(), tableName.getTableName(), filter).orElseThrow(() -> new TableNotFoundException(tableName));
}

Example 29 with PrestoException

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

the class BackgroundHiveSplitLoader method listAndSortBucketFiles.

private static List<LocatedFileStatus> listAndSortBucketFiles(HiveFileIterator hiveFileIterator, int bucketCount) {
    ArrayList<LocatedFileStatus> list = new ArrayList<>(bucketCount);
    while (hiveFileIterator.hasNext()) {
        LocatedFileStatus next = hiveFileIterator.next();
        if (isDirectory(next)) {
            // Fail here to be on the safe side. This seems to be the same as what Hive does
            throw new PrestoException(HIVE_INVALID_BUCKET_FILES, format("%s Found sub-directory in bucket directory for partition: %s", CORRUPT_BUCKETING, hiveFileIterator.getPartitionName()));
        }
        list.add(next);
    }
    if (list.size() != bucketCount) {
        throw new PrestoException(HIVE_INVALID_BUCKET_FILES, format("%s The number of files in the directory (%s) does not match the declared bucket count (%s) for partition: %s", CORRUPT_BUCKETING, list.size(), bucketCount, hiveFileIterator.getPartitionName()));
    }
    // Sort FileStatus objects (instead of, e.g., fileStatus.getPath().toString). This matches org.apache.hadoop.hive.ql.metadata.Table.getSortedPaths
    list.sort(null);
    return list;
}

Example 30 with PrestoException

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

the class BackgroundHiveSplitLoader method loadPartition.

private void loadPartition(HivePartitionMetadata partition) throws IOException {
    String partitionName = partition.getHivePartition().getPartitionId();
    Properties schema = getPartitionSchema(table, partition.getPartition());
    List<HivePartitionKey> partitionKeys = getPartitionKeys(table, partition.getPartition());
    TupleDomain<HiveColumnHandle> effectivePredicate = partition.getHivePartition().getEffectivePredicate();
    Path path = new Path(getPartitionLocation(table, partition.getPartition()));
    Configuration configuration = hdfsEnvironment.getConfiguration(path);
    InputFormat<?, ?> inputFormat = getInputFormat(configuration, schema, false);
    FileSystem fs = hdfsEnvironment.getFileSystem(session.getUser(), path);
    if (inputFormat instanceof SymlinkTextInputFormat) {
        if (bucketHandle.isPresent()) {
            throw new PrestoException(StandardErrorCode.NOT_SUPPORTED, "Bucketed table in SymlinkTextInputFormat is not yet supported");
        }
        // TODO: This should use an iterator like the HiveFileIterator
        for (Path targetPath : getTargetPathsFromSymlink(fs, path)) {
            // The input should be in TextInputFormat.
            TextInputFormat targetInputFormat = new TextInputFormat();
            // get the configuration for the target path -- it may be a different hdfs instance
            Configuration targetConfiguration = hdfsEnvironment.getConfiguration(targetPath);
            JobConf targetJob = new JobConf(targetConfiguration);
            targetJob.setInputFormat(TextInputFormat.class);
            targetInputFormat.configure(targetJob);
            FileInputFormat.setInputPaths(targetJob, targetPath);
            InputSplit[] targetSplits = targetInputFormat.getSplits(targetJob, 0);
            if (addSplitsToSource(targetSplits, partitionName, partitionKeys, schema, effectivePredicate, partition.getColumnCoercions())) {
                return;
            }
        }
        return;
    }
    // on the input format to obtain file splits.
    if (shouldUseFileSplitsFromInputFormat(inputFormat)) {
        JobConf jobConf = new JobConf(configuration);
        FileInputFormat.setInputPaths(jobConf, path);
        InputSplit[] splits = inputFormat.getSplits(jobConf, 0);
        addSplitsToSource(splits, partitionName, partitionKeys, schema, effectivePredicate, partition.getColumnCoercions());
        return;
    }
    // If only one bucket could match: load that one file
    HiveFileIterator iterator = new HiveFileIterator(path, fs, directoryLister, namenodeStats, partitionName, inputFormat, schema, partitionKeys, effectivePredicate, partition.getColumnCoercions());
    if (!buckets.isEmpty()) {
        int bucketCount = buckets.get(0).getBucketCount();
        List<LocatedFileStatus> list = listAndSortBucketFiles(iterator, bucketCount);
        List<Iterator<HiveSplit>> iteratorList = new ArrayList<>();
        for (HiveBucket bucket : buckets) {
            int bucketNumber = bucket.getBucketNumber();
            LocatedFileStatus file = list.get(bucketNumber);
            boolean splittable = isSplittable(iterator.getInputFormat(), hdfsEnvironment.getFileSystem(session.getUser(), file.getPath()), file.getPath());
            iteratorList.add(createHiveSplitIterator(iterator.getPartitionName(), file.getPath().toString(), file.getBlockLocations(), 0, file.getLen(), iterator.getSchema(), iterator.getPartitionKeys(), splittable, session, OptionalInt.of(bucketNumber), effectivePredicate, partition.getColumnCoercions()));
        }
        addToHiveSplitSourceRoundRobin(iteratorList);
        return;
    }
    // If table is bucketed: list the directory, sort, tag with bucket id
    if (bucketHandle.isPresent()) {
        // HiveFileIterator skips hidden files automatically.
        int bucketCount = bucketHandle.get().getBucketCount();
        List<LocatedFileStatus> list = listAndSortBucketFiles(iterator, bucketCount);
        List<Iterator<HiveSplit>> iteratorList = new ArrayList<>();
        for (int bucketIndex = 0; bucketIndex < bucketCount; bucketIndex++) {
            LocatedFileStatus file = list.get(bucketIndex);
            boolean splittable = isSplittable(iterator.getInputFormat(), hdfsEnvironment.getFileSystem(session.getUser(), file.getPath()), file.getPath());
            iteratorList.add(createHiveSplitIterator(iterator.getPartitionName(), file.getPath().toString(), file.getBlockLocations(), 0, file.getLen(), iterator.getSchema(), iterator.getPartitionKeys(), splittable, session, OptionalInt.of(bucketIndex), iterator.getEffectivePredicate(), partition.getColumnCoercions()));
        }
        addToHiveSplitSourceRoundRobin(iteratorList);
        return;
    }
    fileIterators.addLast(iterator);
}