vladislav-sidorovich commented on code in PR #15407: URL: https://github.com/apache/iceberg/pull/15407#discussion_r3491111844
########## delta-lake/src/main/java/org/apache/iceberg/delta/BaseSnapshotDeltaLakeKernelTableAction.java: ########## @@ -0,0 +1,629 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.iceberg.delta; + +import io.delta.kernel.Scan; +import io.delta.kernel.Snapshot; +import io.delta.kernel.Table; +import io.delta.kernel.data.ColumnarBatch; +import io.delta.kernel.data.FilteredColumnarBatch; +import io.delta.kernel.data.Row; +import io.delta.kernel.defaults.engine.DefaultEngine; +import io.delta.kernel.engine.Engine; +import io.delta.kernel.exceptions.TableNotFoundException; +import io.delta.kernel.internal.DeltaHistoryManager; +import io.delta.kernel.internal.DeltaLogActionUtils; +import io.delta.kernel.internal.SnapshotImpl; +import io.delta.kernel.internal.TableImpl; +import io.delta.kernel.internal.actions.AddFile; +import io.delta.kernel.internal.actions.RemoveFile; +import io.delta.kernel.internal.fs.Path; +import io.delta.kernel.internal.util.VectorUtils; +import io.delta.kernel.types.StructType; +import io.delta.kernel.utils.CloseableIterator; +import java.io.File; +import java.io.IOException; +import java.io.UncheckedIOException; +import java.net.URI; +import java.util.Iterator; +import java.util.List; +import java.util.Map; +import java.util.Optional; +import java.util.Set; +import java.util.function.Supplier; +import java.util.stream.Collectors; +import javax.annotation.Nonnull; +import org.apache.hadoop.conf.Configuration; +import org.apache.iceberg.AppendFiles; +import org.apache.iceberg.DataFile; +import org.apache.iceberg.DataFiles; +import org.apache.iceberg.DeleteFile; +import org.apache.iceberg.DeleteFiles; +import org.apache.iceberg.FileFormat; +import org.apache.iceberg.ManageSnapshots; +import org.apache.iceberg.Metrics; +import org.apache.iceberg.MetricsConfig; +import org.apache.iceberg.OverwriteFiles; +import org.apache.iceberg.PartitionField; +import org.apache.iceberg.PartitionSpec; +import org.apache.iceberg.RowDelta; +import org.apache.iceberg.Schema; +import org.apache.iceberg.TableProperties; +import org.apache.iceberg.Transaction; +import org.apache.iceberg.catalog.Catalog; +import org.apache.iceberg.catalog.TableIdentifier; +import org.apache.iceberg.deletes.BaseDVFileWriter; +import org.apache.iceberg.deletes.DVFileWriter; +import org.apache.iceberg.exceptions.NotFoundException; +import org.apache.iceberg.hadoop.HadoopFileIO; +import org.apache.iceberg.io.InputFile; +import org.apache.iceberg.io.OutputFileFactory; +import org.apache.iceberg.mapping.MappingUtil; +import org.apache.iceberg.mapping.NameMapping; +import org.apache.iceberg.mapping.NameMappingParser; +import org.apache.iceberg.parquet.ParquetUtil; +import org.apache.iceberg.relocated.com.google.common.annotations.VisibleForTesting; +import org.apache.iceberg.relocated.com.google.common.base.Preconditions; +import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap; +import org.apache.iceberg.relocated.com.google.common.collect.Iterators; +import org.apache.iceberg.relocated.com.google.common.collect.Lists; +import org.apache.iceberg.relocated.com.google.common.collect.Sets; +import org.apache.iceberg.types.Type; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +class BaseSnapshotDeltaLakeKernelTableAction implements SnapshotDeltaLakeTable { + private static final Logger LOG = + LoggerFactory.getLogger(BaseSnapshotDeltaLakeKernelTableAction.class); + + private static final String SNAPSHOT_SOURCE_PROP = "snapshot_source"; + private static final String DELTA_SOURCE_VALUE = "delta"; + private static final String ORIGINAL_LOCATION_PROP = "original_location"; + private static final String DELTA_VERSION_TAG_PREFIX = "delta-version-"; + private static final String DELTA_TIMESTAMP_TAG_PREFIX = "delta-ts-"; + private static final Set<String> UNSUPPORTED_DELTA_OPERATIONS = + Set.of( + "ADD COLUMNS", + "CHANGE COLUMN", + "RENAME COLUMN", + "DROP COLUMN", + "ADD CONSTRAINT", + "SET TBLPROPERTIES"); // The operations will be supported eventually + + private final ImmutableMap.Builder<String, String> icebergPropertiesBuilder = + ImmutableMap.builder(); + + private final String deltaTableLocation; + private Engine deltaEngine; + private TableImpl deltaTable; + private SnapshotImpl deltaLatestSnapshot; + + private Catalog icebergCatalog; + private TableIdentifier newTableIdentifier; + private String newTableLocation; + private HadoopFileIO deltaLakeFileIO; + private DeletionVectorConverter deletionVectorConverter; + private OutputFileFactory icebergDVFileFactory; + private final Set<Long> deltaTimestampTags = Sets.newHashSet(); + + BaseSnapshotDeltaLakeKernelTableAction(String deltaTableLocation) { + this.deltaTableLocation = deltaTableLocation; + this.newTableLocation = deltaTableLocation; + } + + @Override + public SnapshotDeltaLakeTable tableProperties(Map<String, String> properties) { + icebergPropertiesBuilder.putAll(properties); + return this; + } + + @Override + public SnapshotDeltaLakeTable tableProperty(String name, String value) { + icebergPropertiesBuilder.put(name, value); + return this; + } + + @Override + public SnapshotDeltaLakeTable tableLocation(String location) { + newTableLocation = location; + return this; + } + + @Override + public SnapshotDeltaLakeTable as(TableIdentifier identifier) { + newTableIdentifier = identifier; + return this; + } + + @Override + public SnapshotDeltaLakeTable icebergCatalog(Catalog catalog) { + icebergCatalog = catalog; + return this; + } + + @Override + public SnapshotDeltaLakeTable deltaLakeConfiguration(Configuration conf) { + deltaEngine = DefaultEngine.create(conf); + deltaLakeFileIO = new HadoopFileIO(conf); + deltaTable = (TableImpl) Table.forPath(deltaEngine, deltaTableLocation); + deletionVectorConverter = new DeletionVectorConverter(deltaEngine, deltaTableLocation); + return this; + } + + @Override + public Result execute() { + Preconditions.checkArgument( + icebergCatalog != null && newTableIdentifier != null, + "Iceberg catalog and identifier cannot be null. Make sure to configure the action with a valid Iceberg catalog and identifier."); + Preconditions.checkArgument( + deltaTable != null && deltaLakeFileIO != null, + "Make sure to configure the action with a valid deltaLakeConfiguration"); + + loadLatestDeltaSnapshot(); + assertDeltaColumnMappingDisabled( + "Conversion of Delta Lake tables with columnMapping feature is not supported."); + + final long latestDeltaVersion = deltaLatestSnapshot.getVersion(); + final long minimalAvailableDeltaVersion = getEarliestRecreatableDeltaLog(); + + SnapshotImpl initialDeltaSnapshot = getDeltaSnapshotAsOfVersion(minimalAvailableDeltaVersion); + + LOG.info( + "Converting Delta Lake table at {} from version {} to version {} into Iceberg table {} ...", + deltaTableLocation, + minimalAvailableDeltaVersion, + latestDeltaVersion, + newTableIdentifier); + + Schema icebergSchema = convertToIcebergSchema(initialDeltaSnapshot.getSchema()); + PartitionSpec partitionSpec = + buildPartitionSpec(icebergSchema, initialDeltaSnapshot.getPartitionColumnNames()); + + Transaction transaction = + icebergCatalog.newCreateTableTransaction( + newTableIdentifier, + icebergSchema, + partitionSpec, + newTableLocation, + buildTablePropertiesWithDelta(initialDeltaSnapshot, deltaTableLocation)); + setDefaultNamingMapping(transaction); + // Create an initial empty snapshot so currentSnapshot() is never null + transaction.newAppend().commit(); + + icebergDVFileFactory = + OutputFileFactory.builderFor(transaction.table(), 1, 1).format(FileFormat.PUFFIN).build(); + + Set<String> processedDataFiles = Sets.newHashSet(); + try { + commitDeltaSnapshotToIcebergTransaction( + initialDeltaSnapshot, transaction, processedDataFiles); + convertEachDeltaVersion( + minimalAvailableDeltaVersion + 1, latestDeltaVersion, transaction, processedDataFiles); + } catch (IOException e) { + throw new UncheckedIOException(e); + } + + transaction.commitTransaction(); + + LOG.info( + "Successfully created Iceberg table {} from Delta Lake table at {}, processed data file count: {}", + newTableIdentifier, + deltaTableLocation, + processedDataFiles.size()); + return processedDataFiles::size; + } + + private void commitDeltaSnapshotToIcebergTransaction( + SnapshotImpl deltaSnapshot, Transaction transaction, Set<String> processedDataFiles) + throws IOException { + Scan scan = deltaSnapshot.getScanBuilder().build(); + try (CloseableIterator<FilteredColumnarBatch> changes = scan.getScanFiles(deltaEngine)) { + + commitDeltaColumnarBatchToIcebergTransaction( + deltaSnapshot.getVersion(), + Iterators.transform(changes, batch -> batch::getData), + transaction, + processedDataFiles); + tagCurrentSnapshot( + deltaSnapshot.getVersion(), deltaSnapshot.getTimestamp(deltaEngine), transaction); + } + } + + private void convertEachDeltaVersion( + long initialDeltaVersion, + long latestDeltaVersion, + Transaction transaction, + Set<String> processedDataFiles) + throws IOException { + + for (long currDeltaVersion = initialDeltaVersion; + currDeltaVersion <= latestDeltaVersion; + currDeltaVersion++) { + try (CloseableIterator<ColumnarBatch> changes = + deltaTable.getChanges( + deltaEngine, + currDeltaVersion, + currDeltaVersion, + Set.of(DeltaLogActionUtils.DeltaAction.values()))) { + + Long commitTimestamp = + commitDeltaColumnarBatchToIcebergTransaction( + currDeltaVersion, + Iterators.transform(changes, batch -> () -> batch), + transaction, + processedDataFiles); + tagCurrentSnapshot(currDeltaVersion, commitTimestamp, transaction); + } + } + } + + /** + * Current implementation uses the schema conversion mapping based on the logical names. Delta + * Lake supports three column mapping modes: none, name, id. So, the renames with columnMapping + * feature can lead to data corruption. + */ + private void assertDeltaColumnMappingDisabled(String errorMessage) { + Map<String, String> configuration = deltaLatestSnapshot.getMetadata().getConfiguration(); + String columnMappingMode = configuration.getOrDefault("delta.columnMapping.mode", "none"); + if (!"none".equals(columnMappingMode)) { + throw new UnsupportedOperationException(errorMessage); + } + } + + private static void setDefaultNamingMapping(Transaction transaction) { + transaction + .table() + .updateProperties() + .set( + TableProperties.DEFAULT_NAME_MAPPING, + NameMappingParser.toJson(MappingUtil.create(transaction.table().schema()))) + .commit(); + } + + private SnapshotImpl getDeltaSnapshotAsOfVersion(long deltaVersion) { + Snapshot snapshot = + Preconditions.checkNotNull( + deltaTable.getSnapshotAsOfVersion(deltaEngine, deltaVersion), + "Delta snapshot for version %s is unreachable.", + deltaVersion); + assertSnapshotImpl(snapshot); + return (SnapshotImpl) snapshot; + } + + /** + * Convert each delta log {@code ColumnarBatch} to Iceberg action and commit to the given {@code + * Transaction}. The complete <a + * href="https://github.com/delta-io/delta/blob/master/PROTOCOL.md#Actions";>spec</a> of delta + * actions. <br> + * Supported: + * <li>Add + * + * @return number of added data files + */ + private Long commitDeltaColumnarBatchToIcebergTransaction( + Long deltaVersion, + Iterator<Supplier<ColumnarBatch>> changes, + Transaction transaction, + Set<String> processedDataFiles) + throws IOException { + + Long originalCommitTimestamp = null; + List<DataFile> dataFilesToAdd = Lists.newArrayList(); + List<DataFile> dataFilesToRemove = Lists.newArrayList(); + List<DeleteFile> deleteFilesToAdd = Lists.newArrayList(); + + while (changes.hasNext()) { + Supplier<ColumnarBatch> container = changes.next(); + ColumnarBatch columnarBatch = container.get(); + try (CloseableIterator<Row> rows = columnarBatch.getRows()) { + while (rows.hasNext()) { + Row row = rows.next(); + if (DeltaLakeActionsTranslationUtil.isCommitInfo(row)) { + Row commitInfo = row.getStruct(row.getSchema().indexOf("commitInfo")); + originalCommitTimestamp = + commitInfo.getLong(commitInfo.getSchema().indexOf("timestamp")); + + assertSupportedDeltaOperation(deltaVersion, commitInfo); + } else if (DeltaLakeActionsTranslationUtil.isAdd(row)) { + AddFile addFile = DeltaLakeActionsTranslationUtil.toAdd(row); + + DataFile dataFile = buildDataFileFromAddDeltaAction(addFile, transaction); + dataFilesToAdd.add(dataFile); + + List<DeleteFile> deleteFiles = + convertDeltaDVsToIcebergDVs(transaction.table().spec(), addFile, dataFile); + + deleteFilesToAdd.addAll(deleteFiles); + processedDataFiles.add(dataFile.location()); + } else if (DeltaLakeActionsTranslationUtil.isRemove(row)) { + RemoveFile remove = DeltaLakeActionsTranslationUtil.toRemove(row); + + DataFile dataFile = buildDataFileFromRemoveDeltaAction(remove, transaction); + dataFilesToRemove.add(dataFile); + processedDataFiles.add(dataFile.location()); + } + } + } + } + + // TODO support more actions + commitIcebergTransaction(transaction, dataFilesToAdd, dataFilesToRemove, deleteFilesToAdd); + + return originalCommitTimestamp; + } + + private List<DeleteFile> convertDeltaDVsToIcebergDVs( + PartitionSpec partitionSpec, AddFile addFile, DataFile dataFile) throws IOException { + if (addFile.getDeletionVector().isEmpty()) { + return List.of(); + } + + DVFileWriter dvWriter = new BaseDVFileWriter(icebergDVFileFactory, path -> null); + try (DVFileWriter closeableWriter = dvWriter) { + long[] positions = + deletionVectorConverter.readDeltaDVPositions(addFile.getDeletionVector().get()); + for (long deletedRowIndex : positions) { + closeableWriter.delete( + dataFile.location(), deletedRowIndex, partitionSpec, dataFile.partition()); + } + } + + return dvWriter.result().deleteFiles(); + } + + /** + * CASES: + * + * <ol> + * <li>Append only + * <li>Delete only + * <li>Append and Delete => overwrite + * <li>RowDelta with deletes. + * <li>No Append, No Delete => No data changes, append tag or snapshot. + * </ol> + */ + private static void commitIcebergTransaction( + Transaction transaction, + List<DataFile> dataFilesToAdd, + List<DataFile> dataFilesToRemove, + List<DeleteFile> deleteFilesToAdd) { + if (!deleteFilesToAdd.isEmpty()) { + // Row Delta + RowDelta rowDelta = transaction.newRowDelta(); + // Avoid validation for multiple DVs added in transaction + // org/apache/iceberg/MergingSnapshotProducer.java:854 + // since we do the conversion sequentially in a single Iceberg transaction + rowDelta.validateFromSnapshot(transaction.table().currentSnapshot().snapshotId()); + + dataFilesToAdd.forEach(rowDelta::addRows); + dataFilesToRemove.forEach(rowDelta::removeRows); + deleteFilesToAdd.forEach(rowDelta::addDeletes); + rowDelta.commit(); + } else if (!dataFilesToAdd.isEmpty() && dataFilesToRemove.isEmpty()) { + // Append only + AppendFiles appendFiles = transaction.newAppend(); + dataFilesToAdd.forEach(appendFiles::appendFile); + appendFiles.commit(); + } else if (dataFilesToAdd.isEmpty() && !dataFilesToRemove.isEmpty()) { + // Delete only + DeleteFiles deleteFiles = transaction.newDelete(); + dataFilesToRemove.forEach(deleteFiles::deleteFile); + deleteFiles.commit(); + } else if (!dataFilesToAdd.isEmpty() && !dataFilesToRemove.isEmpty()) { + // Overwrite + OverwriteFiles overwriteFiles = transaction.newOverwrite(); + dataFilesToAdd.forEach(overwriteFiles::addFile); + dataFilesToRemove.forEach(overwriteFiles::deleteFile); + overwriteFiles.commit(); + } else { + // Tag case + transaction.newAppend().commit(); + } + } + + private DataFile buildDataFileFromAddDeltaAction(AddFile addFile, Transaction transaction) { + String path = addFile.getPath(); + long dataFileSize = addFile.getSize(); + String fullFilePath = getFullFilePath(path, deltaTable.getPath(deltaEngine)); + + InputFile inputDataFile = deltaLakeFileIO.newInputFile(fullFilePath); + if (!inputDataFile.exists()) { + throw new NotFoundException( + "File %s is referenced in the logs of Delta Lake table at %s, but cannot be found in the storage", + fullFilePath, deltaTableLocation); + } + + MetricsConfig metricsConfig = MetricsConfig.forTable(transaction.table()); + String nameMappingString = + transaction.table().properties().get(TableProperties.DEFAULT_NAME_MAPPING); + NameMapping nameMapping = + nameMappingString != null ? NameMappingParser.fromJson(nameMappingString) : null; + // TODO read metrics from Delta log to avoid data flies read + Metrics metrics = ParquetUtil.fileMetrics(inputDataFile, metricsConfig, nameMapping); + + Map<String, String> partitionValues = VectorUtils.toJavaMap(addFile.getPartitionValues()); + PartitionSpec partitionSpec = transaction.table().spec(); + List<String> partitionValueList = + partitionSpec.fields().stream() + .map(PartitionField::name) + .map(partitionValues::get) + .collect(Collectors.toList()); + + return DataFiles.builder(partitionSpec) + .withPath(fullFilePath) + .withFormat(FileFormat.PARQUET) // Delta supports only parquet datafiles + .withFileSizeInBytes(dataFileSize) + .withMetrics(metrics) + .withPartitionValues(partitionValueList) + .build(); + } + + private DataFile buildDataFileFromRemoveDeltaAction( + RemoveFile removeFile, Transaction transaction) { + String path = removeFile.getPath(); + String fullFilePath = getFullFilePath(path, deltaTable.getPath(deltaEngine)); + + InputFile inputDataFile = deltaLakeFileIO.newInputFile(fullFilePath); + if (!inputDataFile.exists()) { Review Comment: I think there is a misunderstanding here. 1. `VACUUM` is fully supported and covered by tests. 2. This check verifies that data file exists not because it was deleted by `VACUUM` but in general verify integrity of the current table snapshot. Also, this data file used below in `ParquetUtil.fileMetrics`, so it's prevention of unclear not-found error. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
