Jackie-Jiang commented on a change in pull request #5934: URL: https://github.com/apache/incubator-pinot/pull/5934#discussion_r483328086
########## File path: pinot-core/src/main/java/org/apache/pinot/core/segment/processing/framework/SegmentReducer.java ########## @@ -0,0 +1,151 @@ +/** + * 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.pinot.core.segment.processing.framework; + +import java.io.File; +import java.io.IOException; +import java.util.Iterator; +import org.apache.avro.file.DataFileWriter; +import org.apache.avro.generic.GenericData; +import org.apache.avro.generic.GenericDatumWriter; +import org.apache.pinot.core.segment.processing.collector.Collector; +import org.apache.pinot.core.segment.processing.collector.CollectorFactory; +import org.apache.pinot.core.segment.processing.utils.SegmentProcessorUtils; +import org.apache.pinot.spi.data.Schema; +import org.apache.pinot.spi.data.readers.GenericRow; +import org.apache.pinot.spi.data.readers.RecordReader; +import org.apache.pinot.spi.data.readers.RecordReaderFactory; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + + +/** + * Reducer phase of the SegmentProcessorFramework + * Reads the avro files in the input directory and creates output avro files in the reducer output directory. + * The avro files in the input directory are expected to contain data for only 1 partition + * Performs operations on that partition data as follows: + * - concatenation/rollup of records + * - split + * - TODO: dedup + */ +public class SegmentReducer { + + private static final Logger LOGGER = LoggerFactory.getLogger(SegmentReducer.class); + private static final int MAX_RECORDS_TO_COLLECT = 5_000_000; + + private final File _reducerInputDir; + private final File _reducerOutputDir; + + private final String _reducerId; + private final Schema _pinotSchema; + private final org.apache.avro.Schema _avroSchema; + private final Collector _collector; + private final int _numRecordsPerPart; + + public SegmentReducer(String reducerId, File reducerInputDir, SegmentReducerConfig reducerConfig, + File reducerOutputDir) { + _reducerInputDir = reducerInputDir; + _reducerOutputDir = reducerOutputDir; + + _reducerId = reducerId; + _pinotSchema = reducerConfig.getPinotSchema(); + _avroSchema = SegmentProcessorUtils.convertPinotSchemaToAvroSchema(_pinotSchema); + _collector = CollectorFactory.getCollector(reducerConfig.getCollectorConfig(), _pinotSchema); + _numRecordsPerPart = reducerConfig.getNumRecordsPerPart(); + LOGGER.info("Initialized reducer with id: {}, input dir: {}, output dir: {}, collector: {}, numRecordsPerPart: {}", + _reducerId, _reducerInputDir, _reducerOutputDir, _collector.getClass(), _numRecordsPerPart); + } + + /** + * Reads the avro files in the input directory. + * Performs configured operations and outputs to other avro file(s) in the reducer output directory. + */ + public void reduce() + throws Exception { + + int part = 0; + for (File inputFile : _reducerInputDir.listFiles()) { + + RecordReader avroRecordReader = RecordReaderFactory + .getRecordReaderByClass("org.apache.pinot.plugin.inputformat.avro.AvroRecordReader", inputFile, + _pinotSchema.getColumnNames(), null); + + while (avroRecordReader.hasNext()) { + GenericRow next = avroRecordReader.next(); + + // Aggregations + _collector.collect(next); + + // Exceeded max records allowed to collect. Flush + if (_collector.size() == MAX_RECORDS_TO_COLLECT) { Review comment: I think it is fine if user configures a very low `_numRecordsPerPart` and not get much aggregation. The problem of using the fixed large threshold is that for use cases with lots of columns, we might run out of memory and there is no knob to tune it. In order to get better aggregation, we can potentially do an external sort on the files before collecting the records. Not requires in the first version. ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
