[HUDI-3478] Implement CDC Write in Spark #6697
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,253 @@ | ||
| /* | ||
| * 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.hudi.io; | ||
|
|
||
| import org.apache.avro.Schema; | ||
| import org.apache.avro.generic.GenericData; | ||
| import org.apache.avro.generic.GenericRecord; | ||
| import org.apache.avro.generic.IndexedRecord; | ||
|
|
||
| import org.apache.hadoop.fs.FileSystem; | ||
| import org.apache.hadoop.fs.Path; | ||
|
|
||
| import org.apache.hudi.avro.HoodieAvroUtils; | ||
| import org.apache.hudi.common.fs.FSUtils; | ||
| import org.apache.hudi.common.model.HoodieAvroPayload; | ||
| import org.apache.hudi.common.model.HoodieRecord; | ||
| import org.apache.hudi.common.model.HoodieWriteStat; | ||
| import org.apache.hudi.common.table.HoodieTableConfig; | ||
| import org.apache.hudi.common.table.cdc.HoodieCDCOperation; | ||
| import org.apache.hudi.common.table.cdc.HoodieCDCSupplementalLoggingMode; | ||
| import org.apache.hudi.common.table.cdc.HoodieCDCUtils; | ||
| import org.apache.hudi.common.table.log.AppendResult; | ||
| import org.apache.hudi.common.table.log.HoodieLogFormat; | ||
| import org.apache.hudi.common.table.log.block.HoodieCDCDataBlock; | ||
| import org.apache.hudi.common.table.log.block.HoodieLogBlock; | ||
| import org.apache.hudi.common.util.DefaultSizeEstimator; | ||
| import org.apache.hudi.common.util.Option; | ||
| import org.apache.hudi.common.util.StringUtils; | ||
| import org.apache.hudi.common.util.collection.ExternalSpillableMap; | ||
| import org.apache.hudi.config.HoodieWriteConfig; | ||
| import org.apache.hudi.exception.HoodieException; | ||
| import org.apache.hudi.exception.HoodieIOException; | ||
| import org.apache.hudi.exception.HoodieUpsertException; | ||
|
|
||
| import java.io.Closeable; | ||
| import java.io.IOException; | ||
| import java.util.Collections; | ||
| import java.util.HashMap; | ||
| import java.util.List; | ||
| import java.util.Map; | ||
| import java.util.stream.Collectors; | ||
|
|
||
| /** | ||
| * This class encapsulates all the cdc-writing functions. | ||
| */ | ||
| public class HoodieCDCLogger implements Closeable { | ||
|
|
||
| private final String commitTime; | ||
|
|
||
| private final String keyField; | ||
|
|
||
| private final Schema dataSchema; | ||
|
|
||
| private final boolean populateMetaFields; | ||
|
|
||
| // writer for cdc data | ||
| private final HoodieLogFormat.Writer cdcWriter; | ||
|
|
||
| private final boolean cdcEnabled; | ||
|
|
||
| private final HoodieCDCSupplementalLoggingMode cdcSupplementalLoggingMode; | ||
|
|
||
| private final Schema cdcSchema; | ||
|
|
||
| private final String cdcSchemaString; | ||
|
|
||
| // the cdc data | ||
| private final Map<String, HoodieAvroPayload> cdcData; | ||
|
|
||
| public HoodieCDCLogger( | ||
| String commitTime, | ||
| HoodieWriteConfig config, | ||
| HoodieTableConfig tableConfig, | ||
| Schema schema, | ||
| HoodieLogFormat.Writer cdcWriter, | ||
| long maxInMemorySizeInBytes) { | ||
| try { | ||
| this.commitTime = commitTime; | ||
| this.dataSchema = HoodieAvroUtils.removeMetadataFields(schema); | ||
| this.populateMetaFields = config.populateMetaFields(); | ||
| this.keyField = populateMetaFields ? HoodieRecord.RECORD_KEY_METADATA_FIELD | ||
| : tableConfig.getRecordKeyFieldProp(); | ||
| this.cdcWriter = cdcWriter; | ||
|
|
||
| this.cdcEnabled = config.getBooleanOrDefault(HoodieTableConfig.CDC_ENABLED); | ||
| this.cdcSupplementalLoggingMode = HoodieCDCSupplementalLoggingMode.parse( | ||
|
Contributor
There was a problem hiding this comment. Is the |
||
| config.getStringOrDefault(HoodieTableConfig.CDC_SUPPLEMENTAL_LOGGING_MODE)); | ||
|
|
||
| if (cdcSupplementalLoggingMode.equals(HoodieCDCSupplementalLoggingMode.WITH_BEFORE_AFTER)) { | ||
| this.cdcSchema = HoodieCDCUtils.CDC_SCHEMA; | ||
| this.cdcSchemaString = HoodieCDCUtils.CDC_SCHEMA_STRING; | ||
| } else if (cdcSupplementalLoggingMode.equals(HoodieCDCSupplementalLoggingMode.WITH_BEFORE)) { | ||
| this.cdcSchema = HoodieCDCUtils.CDC_SCHEMA_OP_RECORDKEY_BEFORE; | ||
| this.cdcSchemaString = HoodieCDCUtils.CDC_SCHEMA_OP_RECORDKEY_BEFORE_STRING; | ||
| } else { | ||
| this.cdcSchema = HoodieCDCUtils.CDC_SCHEMA_OP_AND_RECORDKEY; | ||
| this.cdcSchemaString = HoodieCDCUtils.CDC_SCHEMA_OP_AND_RECORDKEY_STRING; | ||
| } | ||
|
|
||
| this.cdcData = new ExternalSpillableMap<>( | ||
|
Contributor
There was a problem hiding this comment. No need to tackle in this PR, but want to call it out that accumulating records in memory will certainly be problematic from memory footprint (as well as GC) perspective. Note, that Spark does process either via iteration where it doesn't hold records (also limited to the micro-batch size if underlying format supports batch reads, otherwise it will hold just 1 record at a time) in memory for longer than the RDD execution chain requires. The only accumulation point is Parquet writer which is much more efficient though: a) relying on encoding as well as compression, b) storing binary/serialized date in memory
Contributor
Author
There was a problem hiding this comment. thank you for pointing this out. i will think deeply about it. |
||
| maxInMemorySizeInBytes, | ||
| config.getSpillableMapBasePath(), | ||
| new DefaultSizeEstimator<>(), | ||
| new DefaultSizeEstimator<>(), | ||
| config.getCommonConfig().getSpillableDiskMapType(), | ||
| config.getCommonConfig().isBitCaskDiskMapCompressionEnabled() | ||
| ); | ||
| } catch (IOException e) { | ||
| throw new HoodieUpsertException("Failed to initialize HoodieCDCLogger", e); | ||
| } | ||
| } | ||
|
|
||
| public void put(HoodieRecord hoodieRecord, | ||
| GenericRecord oldRecord, | ||
| Option<IndexedRecord> newRecord) { | ||
| if (cdcEnabled) { | ||
| String recordKey = hoodieRecord.getRecordKey(); | ||
| GenericData.Record cdcRecord; | ||
| if (newRecord.isPresent()) { | ||
| GenericRecord record = (GenericRecord) newRecord.get(); | ||
| if (oldRecord == null) { | ||
| // inserted cdc record | ||
| cdcRecord = createCDCRecord(HoodieCDCOperation.INSERT, recordKey, | ||
| null, record); | ||
| } else { | ||
| // updated cdc record | ||
| cdcRecord = createCDCRecord(HoodieCDCOperation.UPDATE, recordKey, | ||
| oldRecord, record); | ||
| } | ||
| } else { | ||
| // deleted cdc record | ||
| cdcRecord = createCDCRecord(HoodieCDCOperation.DELETE, recordKey, | ||
| oldRecord, null); | ||
| } | ||
| cdcData.put(recordKey, new HoodieAvroPayload(Option.of(cdcRecord))); | ||
| } | ||
| } | ||
|
|
||
| private GenericData.Record createCDCRecord(HoodieCDCOperation operation, | ||
| String recordKey, | ||
| GenericRecord oldRecord, | ||
| GenericRecord newRecord) { | ||
| GenericData.Record record; | ||
| if (cdcSupplementalLoggingMode.equals(HoodieCDCSupplementalLoggingMode.WITH_BEFORE_AFTER)) { | ||
| record = HoodieCDCUtils.cdcRecord(operation.getValue(), commitTime, | ||
| removeCommitMetadata(oldRecord), newRecord); | ||
| } else if (cdcSupplementalLoggingMode.equals(HoodieCDCSupplementalLoggingMode.WITH_BEFORE)) { | ||
| record = HoodieCDCUtils.cdcRecord(operation.getValue(), recordKey, | ||
| removeCommitMetadata(oldRecord)); | ||
| } else { | ||
| record = HoodieCDCUtils.cdcRecord(operation.getValue(), recordKey); | ||
| } | ||
| return record; | ||
| } | ||
|
|
||
| private GenericRecord removeCommitMetadata(GenericRecord record) { | ||
|
Contributor
There was a problem hiding this comment. We can re-use |
||
| return HoodieAvroUtils.rewriteRecordWithNewSchema(record, dataSchema, new HashMap<>()); | ||
| } | ||
|
|
||
| public boolean isEmpty() { | ||
| return !this.cdcEnabled || this.cdcData.isEmpty(); | ||
| } | ||
|
|
||
| public Option<AppendResult> writeCDCData() { | ||
| if (isEmpty()) { | ||
| return Option.empty(); | ||
| } | ||
|
|
||
| try { | ||
| List<IndexedRecord> records = cdcData.values().stream() | ||
| .map(record -> { | ||
| try { | ||
| return record.getInsertValue(cdcSchema).get(); | ||
| } catch (IOException e) { | ||
| throw new HoodieIOException("Failed to get cdc record", e); | ||
| } | ||
| }).collect(Collectors.toList()); | ||
|
|
||
| Map<HoodieLogBlock.HeaderMetadataType, String> header = new HashMap<>(); | ||
| header.put(HoodieLogBlock.HeaderMetadataType.INSTANT_TIME, commitTime); | ||
| header.put(HoodieLogBlock.HeaderMetadataType.SCHEMA, cdcSchemaString); | ||
|
|
||
| HoodieLogBlock block = new HoodieCDCDataBlock(records, header, keyField); | ||
| AppendResult result = cdcWriter.appendBlocks(Collections.singletonList(block)); | ||
|
|
||
| // call close to trigger the data flush. | ||
| this.close(); | ||
|
|
||
| return Option.of(result); | ||
| } catch (Exception e) { | ||
| throw new HoodieException("Failed to write the cdc data to " + cdcWriter.getLogFile().getPath(), e); | ||
| } | ||
| } | ||
|
|
||
| @Override | ||
| public void close() { | ||
| try { | ||
| if (cdcWriter != null) { | ||
| cdcWriter.close(); | ||
| } | ||
| } catch (IOException e) { | ||
| throw new HoodieIOException("Failed to close HoodieCDCLogger", e); | ||
| } finally { | ||
| cdcData.clear(); | ||
| } | ||
| } | ||
|
|
||
| public static Option<AppendResult> writeCDCDataIfNeeded(HoodieCDCLogger cdcLogger, | ||
|
Contributor
There was a problem hiding this comment. This method we can move back to |
||
| long recordsWritten, | ||
| long insertRecordsWritten) { | ||
| if (cdcLogger == null || recordsWritten == 0L || (recordsWritten == insertRecordsWritten)) { | ||
| // the following cases where we do not need to write out the cdc file: | ||
| // case 1: all the data from the previous file slice are deleted. and no new data is inserted; | ||
| // case 2: all the data are new-coming, | ||
| return Option.empty(); | ||
| } | ||
| return cdcLogger.writeCDCData(); | ||
| } | ||
|
|
||
| public static void setCDCStatIfNeeded(HoodieWriteStat stat, | ||
| Option<AppendResult> cdcResult, | ||
| String partitionPath, | ||
| FileSystem fs) { | ||
| try { | ||
| if (cdcResult.isPresent()) { | ||
| Path cdcLogFile = cdcResult.get().logFile().getPath(); | ||
| String cdcFileName = cdcLogFile.getName(); | ||
| String cdcPath = StringUtils.isNullOrEmpty(partitionPath) ? cdcFileName : partitionPath + "/" + cdcFileName; | ||
| long cdcFileSizeInBytes = FSUtils.getFileSize(fs, cdcLogFile); | ||
| stat.setCdcPath(cdcPath); | ||
| stat.setCdcWriteBytes(cdcFileSizeInBytes); | ||
| } | ||
| } catch (IOException e) { | ||
| throw new HoodieUpsertException("Failed to set cdc write stat", e); | ||
| } | ||
| } | ||
| } | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.