rdblue commented on a change in pull request #25354: [SPARK-28612][SQL] Add DataFrameWriterV2 API URL: https://github.com/apache/spark/pull/25354#discussion_r316935230
########## File path: sql/catalyst/src/main/java/org/apache/spark/sql/catalyst/expressions/PartitionTransforms.scala ########## @@ -0,0 +1,77 @@ +/* + * 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.catalyst.expressions + +import org.apache.spark.sql.types.{DataType, IntegerType} + +/** + * Base class for expressions that are converted to v2 partition transforms. + * + * Subclasses represent abstract transform functions with concrete implementations that are + * determined by data source implementations. Because the concrete implementation is not known, + * these expressions are [[Unevaluable]]. + * + * These expressions are used to pass transformations from the DataFrame API: + * + * {{{ + * df.writeTo("catalog.db.table").partitionedBy($"category", days($"timestamp")).create() + * }}} + */ +abstract class PartitionTransformExpression extends Expression with Unevaluable { + override def nullable: Boolean = true +} + +/** + * Expression for the v2 partition transform years. + */ +case class Years(child: Expression) extends PartitionTransformExpression { + override def dataType: DataType = IntegerType + override def children: Seq[Expression] = Seq(child) +} + +/** + * Expression for the v2 partition transform months. + */ +case class Months(child: Expression) extends PartitionTransformExpression { + override def dataType: DataType = IntegerType + override def children: Seq[Expression] = Seq(child) +} + +/** + * Expression for the v2 partition transform days. + */ +case class Days(child: Expression) extends PartitionTransformExpression { + override def dataType: DataType = IntegerType + override def children: Seq[Expression] = Seq(child) +} + +/** + * Expression for the v2 partition transform hours. + */ +case class Hours(child: Expression) extends PartitionTransformExpression { + override def dataType: DataType = IntegerType + override def children: Seq[Expression] = Seq(child) +} + +/** + * Expression for the v2 partition transform bucket. + */ +case class Bucket(numBuckets: Literal, child: Expression) extends PartitionTransformExpression { Review comment: Unfortunately, Spark can't force sources to use a specific bucket function. A great example is Hive tables, which use a Hive-specific function. So the partitioning transforms are general. It's up to the source what function to use when bucketing. When we add support for bucketed joins, for example, we will need a way to look up a UDF with the source's bucketing function implementation. The FunctionCatalog interface I proposed does that. Spark would look up the "bucket" function in the table's catalog to get a concrete implementation that it can use to partition the other side of the bucketed join. If that side is also bucketed, we will want a way to compare the bucket functions to see if they are the same. `equals` will probably work. ---------------------------------------------------------------- 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] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
