rdblue commented on code in PR #5872: URL: https://github.com/apache/iceberg/pull/5872#discussion_r990822073
########## spark/v3.3/spark/src/main/java/org/apache/iceberg/spark/source/SparkPushedDownAggregateUtil.java: ########## @@ -0,0 +1,376 @@ +/* + * 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.spark.source; + +import java.math.BigDecimal; +import java.nio.ByteBuffer; +import java.util.List; +import java.util.Locale; +import java.util.Map; +import org.apache.iceberg.MetadataTableType; +import org.apache.iceberg.Schema; +import org.apache.iceberg.Table; +import org.apache.iceberg.expressions.Expression; +import org.apache.iceberg.expressions.Literal; +import org.apache.iceberg.expressions.UnboundAggregate; +import org.apache.iceberg.relocated.com.google.common.collect.Lists; +import org.apache.iceberg.spark.SparkSchemaUtil; +import org.apache.iceberg.spark.SparkTableUtil; +import org.apache.iceberg.types.Conversions; +import org.apache.iceberg.types.Type; +import org.apache.iceberg.types.Types; +import org.apache.spark.sql.Dataset; +import org.apache.spark.sql.Row; +import org.apache.spark.sql.SparkSession; +import org.apache.spark.sql.catalyst.InternalRow; +import org.apache.spark.sql.types.ArrayType; +import org.apache.spark.sql.types.DataType; +import org.apache.spark.sql.types.DataTypes; +import org.apache.spark.sql.types.Decimal; +import org.apache.spark.sql.types.DecimalType; +import org.apache.spark.sql.types.MapType; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import scala.collection.JavaConverters; + +/** Helper methods for working with Spark aggregate push down. */ +public class SparkPushedDownAggregateUtil { + + private SparkPushedDownAggregateUtil() {} + + // Build schema for pushed down aggregates. This schema will be used as the scan schema. + public static StructType buildSchemaForPushedDownAggregate( + List<Expression> aggregates, boolean caseSensitive, Schema schema) { + StructType finalSchema = new StructType(); + for (int index = 0; index < aggregates.size(); index++) { + if ((aggregates.get(index)).op().name().equals("COUNTSTAR")) { + StructField field = + new StructField("COUNT(*)", DataTypes.LongType, false, Metadata.empty()); + finalSchema = finalSchema.add(field); + } else { + String colName = ((UnboundAggregate) aggregates.get(index)).ref().name(); + DataType dataType = getDataTypeForAggregateColumn(colName, caseSensitive, schema); + if (dataType instanceof StructType + || dataType instanceof ArrayType + || dataType instanceof MapType) { + // not building pushed down aggregate schema for complex types to disable aggregate push + // down because the statistic info for complex are not available. + return finalSchema; + } + if ((aggregates.get(index)).op().name().equals("COUNT")) { + StructField field = + new StructField( + "COUNT(" + colName + ")", DataTypes.LongType, false, Metadata.empty()); + finalSchema = finalSchema.add(field); + } else if ((aggregates.get(index)).op().name().equals("MAX")) { + StructField field = + new StructField("MAX(" + colName + ")", dataType, false, Metadata.empty()); + finalSchema = finalSchema.add(field); + } else if ((aggregates.get(index)).op().name().equals("MIN")) { + StructField field = + new StructField("MIN(" + colName + ")", dataType, false, Metadata.empty()); + finalSchema = finalSchema.add(field); + } + } + } + return finalSchema; + } + + public static DataType getDataTypeForAggregateColumn( + String colName, boolean caseSensitive, Schema schema) { + Type type = null; + for (int i = 0; i < schema.columns().size(); i++) { + if ((caseSensitive && schema.columns().get(i).name().equals(colName)) + || (!caseSensitive && schema.columns().get(i).name().equalsIgnoreCase(colName))) { + type = schema.columns().get(i).type(); + } + } + return SparkSchemaUtil.convert(type); + } + + /** + * get "lower_bounds", "upper_bounds", "record_count", "null_value_counts" from metadata table and + * use these to calculate Max/Min/Count, and then use the values of Max/Min/Count to construct an + * InternalRow to return to Spark. + */ + @SuppressWarnings("checkstyle:CyclomaticComplexity") + public static InternalRow[] constructInternalRowForPushedDownAggregate( + SparkSession spark, Table table, StructType pushedAggregateSchema, boolean caseSensitive) { + // get the statistics info from DATA_FILES, calculate the aggregate values (min/max/count) + // and use these value to build an internalRow. + Dataset<Row> metadataRows = + SparkTableUtil.loadMetadataTable(spark, table, MetadataTableType.DATA_FILES); + + Dataset dataset = + metadataRows.selectExpr( + "lower_bounds", "upper_bounds", "record_count", "null_value_counts"); + Row[] statisticRows = (Row[]) dataset.collect(); + + StructField[] fields = pushedAggregateSchema.fields(); + List<Object> valuesInSparkInternalRow = Lists.newArrayList(); + + for (int i = 0; i < fields.length; i++) { + if (fields[i].name().contains("MIN")) { Review Comment: The field name should not be used for this. It is going to be very slow. Instead, I think this should use the aggregate expressions. Why not implement `eval` for the aggregate expressions and just delegate to that? -- 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]
