spark BatchScanExec 源码

spark BatchScanExec 代码

文件路径：/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/BatchScanExec.scala

/*
 * 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.execution.datasources.v2

import com.google.common.base.Objects

import org.apache.spark.SparkException
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.catalyst.InternalRow
import org.apache.spark.sql.catalyst.expressions._
import org.apache.spark.sql.catalyst.plans.QueryPlan
import org.apache.spark.sql.catalyst.plans.physical.{KeyGroupedPartitioning, SinglePartition}
import org.apache.spark.sql.catalyst.util.InternalRowSet
import org.apache.spark.sql.catalyst.util.truncatedString
import org.apache.spark.sql.connector.catalog.Table
import org.apache.spark.sql.connector.read.{HasPartitionKey, InputPartition, PartitionReaderFactory, Scan, SupportsRuntimeV2Filtering}

/**
 * Physical plan node for scanning a batch of data from a data source v2.
 */
case class BatchScanExec(
    output: Seq[AttributeReference],
    @transient scan: Scan,
    runtimeFilters: Seq[Expression],
    keyGroupedPartitioning: Option[Seq[Expression]] = None,
    ordering: Option[Seq[SortOrder]] = None,
    @transient table: Table) extends DataSourceV2ScanExecBase {

  @transient lazy val batch = scan.toBatch

  // TODO: unify the equal/hashCode implementation for all data source v2 query plans.
  override def equals(other: Any): Boolean = other match {
    case other: BatchScanExec =>
      this.batch == other.batch && this.runtimeFilters == other.runtimeFilters
    case _ =>
      false
  }

  override def hashCode(): Int = Objects.hashCode(batch, runtimeFilters)

  @transient override lazy val inputPartitions: Seq[InputPartition] = batch.planInputPartitions()

  @transient private lazy val filteredPartitions: Seq[Seq[InputPartition]] = {
    val dataSourceFilters = runtimeFilters.flatMap {
      case DynamicPruningExpression(e) => DataSourceV2Strategy.translateRuntimeFilterV2(e)
      case _ => None
    }

    if (dataSourceFilters.nonEmpty) {
      val originalPartitioning = outputPartitioning

      // the cast is safe as runtime filters are only assigned if the scan can be filtered
      val filterableScan = scan.asInstanceOf[SupportsRuntimeV2Filtering]
      filterableScan.filter(dataSourceFilters.toArray)

      // call toBatch again to get filtered partitions
      val newPartitions = scan.toBatch.planInputPartitions()

      originalPartitioning match {
        case p: KeyGroupedPartitioning =>
          if (newPartitions.exists(!_.isInstanceOf[HasPartitionKey])) {
            throw new SparkException("Data source must have preserved the original partitioning " +
                "during runtime filtering: not all partitions implement HasPartitionKey after " +
                "filtering")
          }

          val newRows = new InternalRowSet(p.expressions.map(_.dataType))
          newRows ++= newPartitions.map(_.asInstanceOf[HasPartitionKey].partitionKey())
          val oldRows = p.partitionValuesOpt.get

          if (oldRows.size != newRows.size) {
            throw new SparkException("Data source must have preserved the original partitioning " +
                "during runtime filtering: the number of unique partition values obtained " +
                s"through HasPartitionKey changed: before ${oldRows.size}, after ${newRows.size}")
          }

          if (!oldRows.forall(newRows.contains)) {
            throw new SparkException("Data source must have preserved the original partitioning " +
                "during runtime filtering: the number of unique partition values obtained " +
                s"through HasPartitionKey remain the same but do not exactly match")
          }

          groupPartitions(newPartitions).get.map(_._2)

        case _ =>
          // no validation is needed as the data source did not report any specific partitioning
        newPartitions.map(Seq(_))
      }

    } else {
      partitions
    }
  }

  override lazy val readerFactory: PartitionReaderFactory = batch.createReaderFactory()

  override lazy val inputRDD: RDD[InternalRow] = {
    val rdd = if (filteredPartitions.isEmpty && outputPartitioning == SinglePartition) {
      // return an empty RDD with 1 partition if dynamic filtering removed the only split
      sparkContext.parallelize(Array.empty[InternalRow], 1)
    } else {
      new DataSourceRDD(
        sparkContext, filteredPartitions, readerFactory, supportsColumnar, customMetrics)
    }
    postDriverMetrics()
    rdd
  }

  override def doCanonicalize(): BatchScanExec = {
    this.copy(
      output = output.map(QueryPlan.normalizeExpressions(_, output)),
      runtimeFilters = QueryPlan.normalizePredicates(
        runtimeFilters.filterNot(_ == DynamicPruningExpression(Literal.TrueLiteral)),
        output))
  }

  override def simpleString(maxFields: Int): String = {
    val truncatedOutputString = truncatedString(output, "[", ", ", "]", maxFields)
    val runtimeFiltersString = s"RuntimeFilters: ${runtimeFilters.mkString("[", ",", "]")}"
    val result = s"$nodeName$truncatedOutputString ${scan.description()} $runtimeFiltersString"
    redact(result)
  }

  override def nodeName: String = {
    s"BatchScan ${table.name()}".trim
  }
}

相关信息

spark 源码目录

相关文章

spark AddPartitionExec 源码

spark AlterNamespaceSetPropertiesExec 源码

spark AlterTableExec 源码

spark CacheTableExec 源码

spark ContinuousScanExec 源码

spark CreateIndexExec 源码

spark CreateNamespaceExec 源码

spark CreateTableExec 源码

spark DataSourceRDD 源码

spark DataSourceV2ScanExecBase 源码