vinothchandar commented on a change in pull request #1402: [HUDI-407] Adding Simple Index URL: https://github.com/apache/incubator-hudi/pull/1402#discussion_r392056290
########## File path: hudi-client/src/main/java/org/apache/hudi/index/bloom/HoodieSimpleIndex.java ########## @@ -0,0 +1,263 @@ +/* + * 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.index.bloom; + +import org.apache.hudi.WriteStatus; +import org.apache.hudi.common.model.HoodieDataFile; +import org.apache.hudi.common.model.HoodieKey; +import org.apache.hudi.common.model.HoodieRecord; +import org.apache.hudi.common.model.HoodieRecordLocation; +import org.apache.hudi.common.model.HoodieRecordPayload; +import org.apache.hudi.common.table.timeline.HoodieInstant; +import org.apache.hudi.common.util.Option; +import org.apache.hudi.common.util.ParquetUtils; +import org.apache.hudi.common.util.collection.Pair; +import org.apache.hudi.config.HoodieWriteConfig; +import org.apache.hudi.table.HoodieTable; + +import com.google.common.annotations.VisibleForTesting; +import org.apache.hadoop.fs.Path; +import org.apache.log4j.LogManager; +import org.apache.log4j.Logger; +import org.apache.spark.api.java.JavaPairRDD; +import org.apache.spark.api.java.JavaRDD; +import org.apache.spark.api.java.JavaSparkContext; +import org.apache.spark.api.java.Optional; +import org.apache.spark.api.java.function.PairFunction; +import org.apache.spark.storage.StorageLevel; + +import java.util.ArrayList; +import java.util.List; +import java.util.Map; + +import scala.Tuple2; + +import static java.util.stream.Collectors.toList; + +/** + * A simple index which reads interested fields from parquet and joins with incoming records to find the tagged location + * + * @param <T> + */ +public class HoodieSimpleIndex<T extends HoodieRecordPayload> extends HoodieBloomIndex<T> { + + private static final Logger LOG = LogManager.getLogger(HoodieSimpleIndex.class); + + public HoodieSimpleIndex(HoodieWriteConfig config) { + super(config); + } + + /** + * Returns an RDD mapping each HoodieKey with a partitionPath/fileID which contains it. Option.Empty if the key is not + * found. + * + * @param hoodieKeys keys to lookup + * @param jsc spark context + * @param hoodieTable hoodie table object + */ + @Override + public JavaPairRDD<HoodieKey, Option<Pair<String, String>>> fetchRecordLocation(JavaRDD<HoodieKey> hoodieKeys, + JavaSparkContext jsc, HoodieTable<T> hoodieTable) { + JavaPairRDD<String, String> partitionRecordKeyPairRDD = + hoodieKeys.mapToPair(key -> new Tuple2<>(key.getPartitionPath(), key.getRecordKey())); + + // Lookup indexes for all the partition/recordkey pair + JavaPairRDD<HoodieKey, HoodieRecordLocation> recordKeyLocationRDD = + lookupIndex(partitionRecordKeyPairRDD, jsc, hoodieTable); + + JavaPairRDD<HoodieKey, String> keyHoodieKeyPairRDD = hoodieKeys.mapToPair(key -> new Tuple2<>(key, null)); + + return keyHoodieKeyPairRDD.leftOuterJoin(recordKeyLocationRDD).mapToPair(keyLoc -> { + Option<Pair<String, String>> partitionPathFileidPair; + if (keyLoc._2._2.isPresent()) { + partitionPathFileidPair = Option.of(Pair.of(keyLoc._1().getPartitionPath(), keyLoc._2._2.get().getFileId())); + } else { + partitionPathFileidPair = Option.empty(); + } + return new Tuple2<>(keyLoc._1, partitionPathFileidPair); + }); + } + + @Override + public JavaRDD<HoodieRecord<T>> tagLocation(JavaRDD<HoodieRecord<T>> recordRDD, JavaSparkContext jsc, + HoodieTable<T> hoodieTable) { + + // Step 0: cache the input record RDD + if (config.getBloomIndexUseCaching()) { + recordRDD.persist(config.getBloomIndexInputStorageLevel()); + } + + // Step 1: Extract out thinner JavaPairRDD of (partitionPath, recordKey) Review comment: We orginally did this to avoid shuffling the input data to do the join..against the bloom filter/range information.. I think in this scenario, all we are doing is a single join against all of records in the DFS partitions.. So I suggest, we make it much simpler and lower overhead, by the following approach - cache the `recordRDD` alone - From that extract affectedPartitions - Prep the RDD of records in these partitions as `partitionFileIdRecLocationEntryPairRDD` - Just join `recordRDD` with that.. it may simpify a few things.. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
