rdblue commented on a change in pull request #4060: URL: https://github.com/apache/iceberg/pull/4060#discussion_r804865402
########## File path: flink/v1.14/flink/src/main/java/org/apache/iceberg/flink/source/reader/ArrayPoolDataIteratorBatcher.java ########## @@ -0,0 +1,126 @@ +/* + * 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.flink.source.reader; + +import java.io.IOException; +import java.io.UncheckedIOException; +import org.apache.flink.configuration.Configuration; +import org.apache.flink.connector.base.source.reader.RecordsWithSplitIds; +import org.apache.flink.connector.base.source.reader.SourceReaderOptions; +import org.apache.flink.connector.file.src.util.Pool; +import org.apache.iceberg.flink.FlinkConfigOptions; +import org.apache.iceberg.flink.source.DataIterator; +import org.apache.iceberg.io.CloseableIterator; + +/** + * This implementation stores record batch in array from recyclable pool + */ +class ArrayPoolDataIteratorBatcher<T> implements DataIteratorBatcher<T> { + private final int batchSize; + private final int handoverQueueSize; + private final RecordFactory<T> recordFactory; + + private transient Pool<T[]> pool; + + ArrayPoolDataIteratorBatcher(Configuration config, RecordFactory<T> recordFactory) { + this.batchSize = config.getInteger(FlinkConfigOptions.SOURCE_READER_FETCH_BATCH_RECORD_COUNT); + this.handoverQueueSize = config.getInteger(SourceReaderOptions.ELEMENT_QUEUE_CAPACITY); + this.recordFactory = recordFactory; + } + + @Override + public CloseableIterator<RecordsWithSplitIds<RecordAndPosition<T>>> batch( + String splitId, DataIterator<T> inputIterator) { + // lazily create pool as it is not serializable + if (pool == null) { + pool = createPoolOfBatches(handoverQueueSize); + } + return new ArrayPoolBatchIterator(splitId, inputIterator, pool); + } + + private Pool<T[]> createPoolOfBatches(int numBatches) { + Pool<T[]> poolOfBatches = new Pool<>(numBatches); + for (int batchId = 0; batchId < numBatches; batchId++) { + T[] batch = recordFactory.createBatch(batchSize); + poolOfBatches.add(batch); + } + return poolOfBatches; + } + + private class ArrayPoolBatchIterator implements CloseableIterator<RecordsWithSplitIds<RecordAndPosition<T>>> { + + private final String splitId; + private final DataIterator<T> inputIterator; + private final Pool<T[]> pool; + + ArrayPoolBatchIterator(String splitId, DataIterator<T> inputIterator, Pool<T[]> pool) { + this.splitId = splitId; + this.inputIterator = inputIterator; + this.pool = pool; + } + + @Override + public boolean hasNext() { + return inputIterator.hasNext(); + } + + @Override + public RecordsWithSplitIds<RecordAndPosition<T>> next() { + T[] batch = getCachedEntry(); + int recordCount = 0; + while (inputIterator.hasNext() && recordCount < batchSize) { + // The record produced by inputIterator can be reused like for the RowData case. + // inputIterator.next() can't be called again until the copy is made + // since the record is not consumed immediately. + T nextRecord = inputIterator.next(); + recordFactory.clone(nextRecord, batch, recordCount); + recordCount++; + if (!inputIterator.currentFileHasNext()) { + // break early so that records in the ArrayResultIterator + // have the same fileOffset. + break; + } + } + + if (recordCount == 0) { + return null; + } else { + return ArrayBatchRecords.forRecords(splitId, pool.recycler(), batch, recordCount, + inputIterator.fileOffset(), inputIterator.recordOffset() - recordCount); + } + } + + @Override + public void close() throws IOException { + if (inputIterator != null) { + inputIterator.close(); + } + } + + private T[] getCachedEntry() { + try { + return pool.pollEntry(); + } catch (InterruptedException e) { + Thread.currentThread().interrupt(); + throw new UncheckedIOException(new IOException("Interrupted")); Review comment: If it isn't an IOException, then we shouldn't create one to throw. You could throw a RuntimeException here. -- 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]
