Examples with HoodieData org.apache.hudi.common.data.HoodieData used on opensource projects

Example 1 with HoodieData

use of org.apache.hudi.common.data.HoodieData in project hudi by apache.

the class TestHoodieCompactor method testWriteStatusContentsAfterCompaction.

@Test
public void testWriteStatusContentsAfterCompaction() throws Exception {
    // insert 100 records
    HoodieWriteConfig config = getConfigBuilder().withCompactionConfig(HoodieCompactionConfig.newBuilder().withMaxNumDeltaCommitsBeforeCompaction(1).build()).build();
    try (SparkRDDWriteClient writeClient = getHoodieWriteClient(config)) {
        String newCommitTime = "100";
        writeClient.startCommitWithTime(newCommitTime);
        List<HoodieRecord> records = dataGen.generateInserts(newCommitTime, 100);
        JavaRDD<HoodieRecord> recordsRDD = jsc.parallelize(records, 1);
        writeClient.insert(recordsRDD, newCommitTime).collect();
        // Update all the 100 records
        HoodieTable table = HoodieSparkTable.create(config, context);
        newCommitTime = "101";
        List<HoodieRecord> updatedRecords = dataGen.generateUpdates(newCommitTime, records);
        JavaRDD<HoodieRecord> updatedRecordsRDD = jsc.parallelize(updatedRecords, 1);
        HoodieIndex index = new HoodieBloomIndex(config, SparkHoodieBloomIndexHelper.getInstance());
        JavaRDD<HoodieRecord> updatedTaggedRecordsRDD = tagLocation(index, updatedRecordsRDD, table);
        writeClient.startCommitWithTime(newCommitTime);
        writeClient.upsertPreppedRecords(updatedTaggedRecordsRDD, newCommitTime).collect();
        metaClient.reloadActiveTimeline();
        // Verify that all data file has one log file
        table = HoodieSparkTable.create(config, context);
        for (String partitionPath : dataGen.getPartitionPaths()) {
            List<FileSlice> groupedLogFiles = table.getSliceView().getLatestFileSlices(partitionPath).collect(Collectors.toList());
            for (FileSlice fileSlice : groupedLogFiles) {
                assertEquals(1, fileSlice.getLogFiles().count(), "There should be 1 log file written for every data file");
            }
        }
        // Do a compaction
        table = HoodieSparkTable.create(config, context);
        String compactionInstantTime = "102";
        table.scheduleCompaction(context, compactionInstantTime, Option.empty());
        table.getMetaClient().reloadActiveTimeline();
        HoodieData<WriteStatus> result = (HoodieData<WriteStatus>) table.compact(context, compactionInstantTime).getWriteStatuses();
        // Verify that all partition paths are present in the WriteStatus result
        for (String partitionPath : dataGen.getPartitionPaths()) {
            List<WriteStatus> writeStatuses = result.collectAsList();
            assertTrue(writeStatuses.stream().filter(writeStatus -> writeStatus.getStat().getPartitionPath().contentEquals(partitionPath)).count() > 0);
        }
    }
}

Example 2 with HoodieData

use of org.apache.hudi.common.data.HoodieData in project hudi by apache.

the class HoodieBackedTableMetadataWriter method update.

/**
 * Update from {@code HoodieRollbackMetadata}.
 *
 * @param rollbackMetadata {@code HoodieRollbackMetadata}
 * @param instantTime Timestamp at which the rollback was performed
 */
@Override
public void update(HoodieRollbackMetadata rollbackMetadata, String instantTime) {
    if (enabled && metadata != null) {
        // Is this rollback of an instant that has been synced to the metadata table?
        String rollbackInstant = rollbackMetadata.getCommitsRollback().get(0);
        boolean wasSynced = metadataMetaClient.getActiveTimeline().containsInstant(new HoodieInstant(false, HoodieTimeline.DELTA_COMMIT_ACTION, rollbackInstant));
        if (!wasSynced) {
            // A compaction may have taken place on metadata table which would have included this instant being rolled back.
            // Revisit this logic to relax the compaction fencing : https://issues.apache.org/jira/browse/HUDI-2458
            Option<String> latestCompaction = metadata.getLatestCompactionTime();
            if (latestCompaction.isPresent()) {
                wasSynced = HoodieTimeline.compareTimestamps(rollbackInstant, HoodieTimeline.LESSER_THAN_OR_EQUALS, latestCompaction.get());
            }
        }
        Map<MetadataPartitionType, HoodieData<HoodieRecord>> records = HoodieTableMetadataUtil.convertMetadataToRecords(engineContext, metadataMetaClient.getActiveTimeline(), rollbackMetadata, getRecordsGenerationParams(), instantTime, metadata.getSyncedInstantTime(), wasSynced);
        commit(instantTime, records, false);
    }
}

Example 3 with HoodieData

use of org.apache.hudi.common.data.HoodieData in project hudi by apache.

the class ListBasedHoodieBloomIndexHelper method findMatchingFilesForRecordKeys.

@Override
public HoodiePairData<HoodieKey, HoodieRecordLocation> findMatchingFilesForRecordKeys(HoodieWriteConfig config, HoodieEngineContext context, HoodieTable hoodieTable, HoodiePairData<String, String> partitionRecordKeyPairs, HoodieData<Pair<String, HoodieKey>> fileComparisonPairs, Map<String, List<BloomIndexFileInfo>> partitionToFileInfo, Map<String, Long> recordsPerPartition) {
    List<Pair<String, HoodieKey>> fileComparisonPairList = HoodieList.getList(fileComparisonPairs).stream().sorted(Comparator.comparing(Pair::getLeft)).collect(toList());
    List<HoodieKeyLookupResult> keyLookupResults = new ArrayList<>();
    Iterator<List<HoodieKeyLookupResult>> iterator = new HoodieBaseBloomIndexCheckFunction(hoodieTable, config).apply(fileComparisonPairList.iterator());
    while (iterator.hasNext()) {
        keyLookupResults.addAll(iterator.next());
    }
    keyLookupResults = keyLookupResults.stream().filter(lr -> lr.getMatchingRecordKeys().size() > 0).collect(toList());
    return context.parallelize(keyLookupResults).flatMap(lookupResult -> lookupResult.getMatchingRecordKeys().stream().map(recordKey -> new ImmutablePair<>(lookupResult, recordKey)).iterator()).mapToPair(pair -> {
        HoodieKeyLookupResult lookupResult = pair.getLeft();
        String recordKey = pair.getRight();
        return new ImmutablePair<>(new HoodieKey(recordKey, lookupResult.getPartitionPath()), new HoodieRecordLocation(lookupResult.getBaseInstantTime(), lookupResult.getFileId()));
    });
}

Example 4 with HoodieData

use of org.apache.hudi.common.data.HoodieData in project hudi by apache.

the class MultipleSparkJobExecutionStrategy method runClusteringForGroupAsync.

/**
 * Submit job to execute clustering for the group.
 */
private CompletableFuture<HoodieData<WriteStatus>> runClusteringForGroupAsync(HoodieClusteringGroup clusteringGroup, Map<String, String> strategyParams, boolean preserveHoodieMetadata, String instantTime) {
    return CompletableFuture.supplyAsync(() -> {
        JavaSparkContext jsc = HoodieSparkEngineContext.getSparkContext(getEngineContext());
        HoodieData<HoodieRecord<T>> inputRecords = readRecordsForGroup(jsc, clusteringGroup, instantTime);
        Schema readerSchema = HoodieAvroUtils.addMetadataFields(new Schema.Parser().parse(getWriteConfig().getSchema()));
        List<HoodieFileGroupId> inputFileIds = clusteringGroup.getSlices().stream().map(info -> new HoodieFileGroupId(info.getPartitionPath(), info.getFileId())).collect(Collectors.toList());
        return performClusteringWithRecordsRDD(inputRecords, clusteringGroup.getNumOutputFileGroups(), instantTime, strategyParams, readerSchema, inputFileIds, preserveHoodieMetadata);
    });
}

Example 5 with HoodieData

use of org.apache.hudi.common.data.HoodieData in project hudi by apache.

the class MultipleSparkJobExecutionStrategy method readRecordsForGroupWithLogs.

/**
 * Read records from baseFiles, apply updates and convert to RDD.
 */
private HoodieData<HoodieRecord<T>> readRecordsForGroupWithLogs(JavaSparkContext jsc, List<ClusteringOperation> clusteringOps, String instantTime) {
    HoodieWriteConfig config = getWriteConfig();
    HoodieTable table = getHoodieTable();
    return HoodieJavaRDD.of(jsc.parallelize(clusteringOps, clusteringOps.size()).mapPartitions(clusteringOpsPartition -> {
        List<Iterator<HoodieRecord<T>>> recordIterators = new ArrayList<>();
        clusteringOpsPartition.forEachRemaining(clusteringOp -> {
            long maxMemoryPerCompaction = IOUtils.getMaxMemoryPerCompaction(new SparkTaskContextSupplier(), config);
            LOG.info("MaxMemoryPerCompaction run as part of clustering => " + maxMemoryPerCompaction);
            try {
                Schema readerSchema = HoodieAvroUtils.addMetadataFields(new Schema.Parser().parse(config.getSchema()));
                HoodieMergedLogRecordScanner scanner = HoodieMergedLogRecordScanner.newBuilder().withFileSystem(table.getMetaClient().getFs()).withBasePath(table.getMetaClient().getBasePath()).withLogFilePaths(clusteringOp.getDeltaFilePaths()).withReaderSchema(readerSchema).withLatestInstantTime(instantTime).withMaxMemorySizeInBytes(maxMemoryPerCompaction).withReadBlocksLazily(config.getCompactionLazyBlockReadEnabled()).withReverseReader(config.getCompactionReverseLogReadEnabled()).withBufferSize(config.getMaxDFSStreamBufferSize()).withSpillableMapBasePath(config.getSpillableMapBasePath()).withPartition(clusteringOp.getPartitionPath()).build();
                Option<HoodieFileReader> baseFileReader = StringUtils.isNullOrEmpty(clusteringOp.getDataFilePath()) ? Option.empty() : Option.of(HoodieFileReaderFactory.getFileReader(table.getHadoopConf(), new Path(clusteringOp.getDataFilePath())));
                HoodieTableConfig tableConfig = table.getMetaClient().getTableConfig();
                recordIterators.add(getFileSliceReader(baseFileReader, scanner, readerSchema, tableConfig.getPayloadClass(), tableConfig.getPreCombineField(), tableConfig.populateMetaFields() ? Option.empty() : Option.of(Pair.of(tableConfig.getRecordKeyFieldProp(), tableConfig.getPartitionFieldProp()))));
            } catch (IOException e) {
                throw new HoodieClusteringException("Error reading input data for " + clusteringOp.getDataFilePath() + " and " + clusteringOp.getDeltaFilePaths(), e);
            }
        });
        return new ConcatenatingIterator<>(recordIterators);
    }));
}