revans2 commented on a change in pull request #29067: URL: https://github.com/apache/spark/pull/29067#discussion_r457604604
########## File path: sql/core/src/main/scala/org/apache/spark/sql/columnar/CachedBatchSerializer.scala ########## @@ -0,0 +1,291 @@ +/* + * 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.spark.sql.columnar + +import org.apache.spark.annotation.{DeveloperApi, Since} +import org.apache.spark.internal.Logging +import org.apache.spark.rdd.RDD +import org.apache.spark.sql.catalyst.InternalRow +import org.apache.spark.sql.catalyst.dsl.expressions._ +import org.apache.spark.sql.catalyst.expressions.{And, Attribute, AttributeReference, BindReferences, EqualNullSafe, EqualTo, Expression, GreaterThan, GreaterThanOrEqual, In, IsNotNull, IsNull, Length, LessThan, LessThanOrEqual, Literal, Or, Predicate, StartsWith} +import org.apache.spark.sql.execution.SparkPlan +import org.apache.spark.sql.execution.columnar.{ColumnStatisticsSchema, PartitionStatistics} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types.{AtomicType, BinaryType, StructType} +import org.apache.spark.sql.vectorized.ColumnarBatch + +/** + * Basic interface that all cached batches of data must support. This is primarily to allow + * for metrics to be handled outside of the encoding and decoding steps in a standard way. + */ +@DeveloperApi +@Since("3.1.0") +trait CachedBatch { + def numRows: Int + def sizeInBytes: Long +} + +/** + * Provides APIs for compressing, filtering, and decompressing SQL data that will be + * persisted/cached. + */ +@DeveloperApi +@Since("3.1.0") +trait CachedBatchSerializer extends Serializable { Review comment: Columnar to row transitions are inserted as a part of the `ApplyColumnarRulesAndInsertTransitions` rule in `QueryExecution.preparations`. It looks for `SparkPlan` instances that output columnar data followed by ones that take row based data as input. The rule inserts a `ColumnarToRowExec` transition between them. This transition supports code generation so in normal operation it will then be wrapped in a `WholeStageCodeGenExec`. `ApplyColumnarRulesAndInsertTransitions` assumes that the output of any plan must be row based to conform with all of the other code that already makes that assumption. It should be a minor change to be able to make that optional so `QueryExecution` could provide a new API that would return a `SparkPlan` that might be columnar and might not. I can put together an prototype of how that would look if you are OK with it. ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
