klsince commented on a change in pull request #7299:
URL: https://github.com/apache/pinot/pull/7299#discussion_r694233398
##########
File path:
pinot-plugins/pinot-batch-ingestion/v0_deprecated/pinot-hadoop/src/main/java/org/apache/pinot/hadoop/job/mappers/SegmentCreationMapper.java
##########
@@ -140,13 +140,16 @@ public void setup(Context context)
setFlagForSchemaMismatch();
// Set up segment name generator
- String segmentNameGeneratorType =
_jobConf.get(JobConfigConstants.SEGMENT_NAME_GENERATOR_TYPE,
JobConfigConstants.DEFAULT_SEGMENT_NAME_GENERATOR);
+ String segmentNameGeneratorType =
Review comment:
nit: it looks like many changes are due to reformatting and renaming
files. maybe open a PR just for those changes to make new changes in this PR
easier to review.
##########
File path:
pinot-plugins/pinot-batch-ingestion/v0_deprecated/pinot-ingestion-common/src/main/java/org/apache/pinot/ingestion/jobs/SegmentPreprocessingJob.java
##########
@@ -90,4 +119,131 @@ protected boolean isDataFile(String fileName) {
// TODO: support orc format in the future.
return fileName.endsWith(".avro");
}
+
+ protected void setTableConfigAndSchema()
+ throws IOException {
+ _tableConfig = getTableConfig();
+ _pinotTableSchema = getSchema();
+
+ Preconditions.checkState(_tableConfig != null, "Table config cannot be
null.");
+ Preconditions.checkState(_pinotTableSchema != null, "Schema cannot be
null");
+ }
+
+ protected void fetchPreProcessingOperations() {
+ _preprocessingOperations = new HashSet<>();
+ TableCustomConfig customConfig = _tableConfig.getCustomConfig();
+ if (customConfig != null) {
+ Map<String, String> customConfigMap = customConfig.getCustomConfigs();
+ if (customConfigMap != null && !customConfigMap.isEmpty()) {
+ String preprocessingOperationsString =
+
customConfigMap.getOrDefault(InternalConfigConstants.PREPROCESS_OPERATIONS, "");
+ DataPreprocessingUtils.getOperations(_preprocessingOperations,
preprocessingOperationsString);
+ }
+ }
+ }
+
+ protected void fetchPartitioningConfig() {
+ // Fetch partition info from table config.
+ if
(!_preprocessingOperations.contains(DataPreprocessingUtils.Operation.PARTITION))
{
+ LOGGER.info("Partitioning is disabled.");
+ return;
+ }
+ SegmentPartitionConfig segmentPartitionConfig =
_tableConfig.getIndexingConfig().getSegmentPartitionConfig();
+ if (segmentPartitionConfig != null) {
+ Map<String, ColumnPartitionConfig> columnPartitionMap =
segmentPartitionConfig.getColumnPartitionMap();
+ Preconditions
+ .checkArgument(columnPartitionMap.size() <= 1, "There should be at
most 1 partition setting in the table.");
+ if (columnPartitionMap.size() == 1) {
+ _partitionColumn = columnPartitionMap.keySet().iterator().next();
+ _numPartitions =
segmentPartitionConfig.getNumPartitions(_partitionColumn);
+ _partitionFunction =
segmentPartitionConfig.getFunctionName(_partitionColumn);
+ }
+ } else {
+ LOGGER.info("Segment partition config is null for table: {}",
_tableConfig.getTableName());
+ }
+ }
+
+ protected void fetchSortingConfig() {
+ if
(!_preprocessingOperations.contains(DataPreprocessingUtils.Operation.SORT)) {
+ LOGGER.info("Sorting is disabled.");
+ return;
+ }
+ // Fetch sorting info from table config first.
+ List<String> sortingColumns = new ArrayList<>();
+ List<FieldConfig> fieldConfigs = _tableConfig.getFieldConfigList();
+ if (fieldConfigs != null && !fieldConfigs.isEmpty()) {
+ for (FieldConfig fieldConfig : fieldConfigs) {
+ if (fieldConfig.getIndexType() == FieldConfig.IndexType.SORTED) {
+ sortingColumns.add(fieldConfig.getName());
+ }
+ }
+ }
+ if (!sortingColumns.isEmpty()) {
+ Preconditions.checkArgument(sortingColumns.size() == 1, "There should be
at most 1 sorted column in the table.");
Review comment:
nit: check `size() <= 1, "There should be at most..."` outside the
if-block?
##########
File path:
pinot-plugins/pinot-batch-ingestion/v0_deprecated/pinot-spark/src/main/java/org/apache/pinot/spark/jobs/preprocess/SparkDataPreprocessingPartitioner.java
##########
@@ -0,0 +1,56 @@
+/**
+ * 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.spark.jobs.preprocess;
+
+import java.util.concurrent.atomic.AtomicInteger;
+import org.apache.pinot.segment.spi.partition.PartitionFunction;
+import org.apache.spark.Partitioner;
+
+
+public class SparkDataPreprocessingPartitioner extends Partitioner {
+ private final int _numPartitions;
+ private final PartitionFunction _partitionFunction;
+ private final AtomicInteger _counter = new AtomicInteger(0);
+
+ public SparkDataPreprocessingPartitioner(int numPartitions,
PartitionFunction partitionFunction) {
+ _numPartitions = numPartitions;
+ _partitionFunction = partitionFunction;
+ }
+
+ @Override
+ public int numPartitions() {
+ return _numPartitions;
+ }
+
+ @Override
+ public int getPartition(Object key) {
+ SparkDataPreprocessingJobKey jobKey = (SparkDataPreprocessingJobKey) key;
Review comment:
why not just call generatePartitionId() in this method?
##########
File path:
pinot-plugins/pinot-batch-ingestion/v0_deprecated/pinot-spark/src/main/java/org/apache/pinot/spark/jobs/preprocess/SparkDataPreprocessingHelper.java
##########
@@ -0,0 +1,153 @@
+/**
+ * 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.spark.jobs.preprocess;
+
+import java.util.ArrayList;
+import java.util.Collection;
+import java.util.Comparator;
+import java.util.List;
+import org.apache.hadoop.fs.Path;
+import org.apache.pinot.ingestion.preprocess.DataPreprocessingHelper;
+import
org.apache.pinot.ingestion.preprocess.partitioners.PartitionFunctionFactory;
+import org.apache.pinot.segment.spi.partition.PartitionFunction;
+import org.apache.pinot.spi.config.table.TableConfig;
+import org.apache.pinot.spi.data.FieldSpec;
+import org.apache.pinot.spi.data.Schema;
+import org.apache.spark.api.java.JavaPairRDD;
+import org.apache.spark.api.java.JavaRDD;
+import org.apache.spark.api.java.function.PairFunction;
+import org.apache.spark.sql.Dataset;
+import org.apache.spark.sql.Row;
+import org.apache.spark.sql.SparkSession;
+import org.apache.spark.sql.types.StructField;
+import org.apache.spark.sql.types.StructType;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+import scala.Tuple2;
+import scala.collection.JavaConverters;
+import scala.collection.Seq;
+
+
+public abstract class SparkDataPreprocessingHelper {
+ private static final Logger LOGGER =
LoggerFactory.getLogger(SparkDataPreprocessingHelper.class);
+
+ protected DataPreprocessingHelper _dataPreprocessingHelper;
+
+ public SparkDataPreprocessingHelper(DataPreprocessingHelper
dataPreprocessingHelper) {
+ _dataPreprocessingHelper = dataPreprocessingHelper;
+ }
+
+ public void registerConfigs(TableConfig tableConfig, Schema tableSchema,
String partitionColumn, int numPartitions,
+ String partitionFunction, String sortingColumn, FieldSpec.DataType
sortingColumnType, int numOutputFiles,
+ int maxNumRecordsPerFile) {
+ _dataPreprocessingHelper
+ .registerConfigs(tableConfig, tableSchema, partitionColumn,
numPartitions, partitionFunction, sortingColumn,
+ sortingColumnType, numOutputFiles, maxNumRecordsPerFile);
+ }
+
+ public void setUpAndExecuteJob(SparkSession sparkSession) {
+ // Read data into data frame.
+ Dataset<Row> dataFrame =
+
sparkSession.read().format(getDataFormat()).load(convertPathsToStrings(_dataPreprocessingHelper._inputDataPaths));
+ JavaRDD<Row> javaRDD = dataFrame.javaRDD();
+
+ // Find positions of partition column and sorting column if specified.
+ StructType schema = dataFrame.schema();
+ StructField[] fields = schema.fields();
+ int partitionColumnPosition = -1;
+ int sortingColumnPosition = -1;
+ PartitionFunction partitionFunction = null;
+ for (int i = 0; i <= fields.length; i++) {
+ StructField field = fields[i];
+ if (_dataPreprocessingHelper._partitionColumn != null &&
_dataPreprocessingHelper._partitionColumn
+ .equalsIgnoreCase(field.name())) {
+ partitionColumnPosition = i;
+ partitionFunction = PartitionFunctionFactory
+ .getPartitionFunction(_dataPreprocessingHelper._partitionFunction,
_dataPreprocessingHelper._numPartitions);
+ }
+ if (_dataPreprocessingHelper._sortingColumn != null &&
_dataPreprocessingHelper._sortingColumn
+ .equalsIgnoreCase(field.name())) {
+ sortingColumnPosition = i;
+ }
+ }
+ int numPartitions;
+ if (partitionColumnPosition == -1) {
+ if (_dataPreprocessingHelper._numOutputFiles > 0) {
+ numPartitions = _dataPreprocessingHelper._numOutputFiles;
+ } else {
+ numPartitions = javaRDD.getNumPartitions();
+ }
+ } else {
+ numPartitions = _dataPreprocessingHelper._numPartitions;
+ }
+ final int finalNumPartitions = numPartitions;
+ final int finalPartitionColumnPosition = partitionColumnPosition;
+ final int finalSortingColumnPosition = sortingColumnPosition;
+ LOGGER.info("Number of partitions: " + finalNumPartitions);
+ LOGGER.info("Position of partition column (if specified): " +
finalPartitionColumnPosition);
+ LOGGER.info("Position of sorting column (if specified): " +
finalSortingColumnPosition);
+ SparkDataPreprocessingPartitioner sparkPartitioner =
+ new SparkDataPreprocessingPartitioner(finalNumPartitions,
partitionFunction);
+
+ // Convert to java pair rdd.
+ JavaPairRDD<Object, Row> pairRDD = javaRDD.mapToPair((PairFunction<Row,
Object, Row>) row -> {
+ Object partitionColumnValue = null;
+ Object sortingColumnValue = null;
+
+ if (_dataPreprocessingHelper._partitionColumn != null) {
+ partitionColumnValue = row.get(finalPartitionColumnPosition);
+ }
+ int partitionId =
sparkPartitioner.generatePartitionId(partitionColumnValue);
+ if (_dataPreprocessingHelper._sortingColumn != null) {
+ sortingColumnValue = row.get(finalSortingColumnPosition);
+ }
+ return new Tuple2<>(new SparkDataPreprocessingJobKey(partitionId,
sortingColumnValue), row);
+ });
+
+ // Repartition and sort within partitions.
+ Comparator<Object> comparator = new SparkDataPreprocessingComparator();
+ JavaPairRDD<Object, Row> partitionedSortedPairRDD =
+ pairRDD.repartitionAndSortWithinPartitions(sparkPartitioner,
comparator);
Review comment:
sometimes use coalesce() (then sortWithinPartitions()) can be more
efficient than repartition(). is it feasible to allow coalesce()?
--
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: commits-unsubscr...@pinot.apache.org
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
