spark-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From hvanhovell <...@git.apache.org>
Subject [GitHub] spark pull request #14136: [SPARK-16282][SQL] Implement percentile SQL funct...
Date Thu, 14 Jul 2016 15:12:53 GMT
Github user hvanhovell commented on a diff in the pull request:

    https://github.com/apache/spark/pull/14136#discussion_r70823811
  
    --- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Percentile.scala
---
    @@ -0,0 +1,172 @@
    +/*
    + * 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.aggregate
    +
    +import org.apache.spark.sql.catalyst.InternalRow
    +import org.apache.spark.sql.catalyst.analysis.TypeCheckResult
    +import org.apache.spark.sql.catalyst.expressions._
    +import org.apache.spark.sql.catalyst.util._
    +import org.apache.spark.sql.types._
    +import org.apache.spark.util.collection.OpenHashMap
    +
    +/**
    + * The Percentile aggregate function computes the exact percentile(s) of expr at pc with
range in
    + * [0, 1].
    + * The parameter pc can be a DoubleType or DoubleType array.
    + *
    + * The operator is bound to the slower sort based aggregation path because the number
of elements
    + * and their partial order cannot be determined in advance. Therefore we have to store
all the
    + * elements in memory, and that too many elements can cause GC paused and eventually
OutOfMemory
    + * Errors.
    + */
    +@ExpressionDescription(
    +  usage = """_FUNC_(epxr, pc) - Returns the percentile(s) of expr at pc (range: [0,1]).
pc can be
    +  a double or double array.""")
    +case class Percentile(
    +  child: Expression,
    +  pc: Expression,
    +  mutableAggBufferOffset: Int = 0,
    +  inputAggBufferOffset: Int = 0) extends ImperativeAggregate {
    +
    +  def this(child: Expression, pc: Expression) = {
    +    this(child = child, pc = pc, mutableAggBufferOffset = 0, inputAggBufferOffset = 0)
    +  }
    +
    +  private val percentiles: Seq[Double] = pc match {
    +    case Literal(ar: GenericArrayData, _: ArrayType) =>
    +      ar.asInstanceOf[GenericArrayData].array.map{ d => d.asInstanceOf[Double]}
    +    case _ => Seq.empty
    +  }
    +
    +  override def prettyName: String = "percentile"
    +
    +  override def withNewMutableAggBufferOffset(newMutableAggBufferOffset: Int): ImperativeAggregate
=
    +    copy(mutableAggBufferOffset = newMutableAggBufferOffset)
    +
    +  override def withNewInputAggBufferOffset(newInputAggBufferOffset: Int): ImperativeAggregate
=
    +    copy(inputAggBufferOffset = newInputAggBufferOffset)
    +
    +  private var counts = new OpenHashMap[Double, Long]()
    +
    +  override def children: Seq[Expression] = child :: pc :: Nil
    +
    +  override def nullable: Boolean = false
    +
    +  override def dataType: DataType = ArrayType(DoubleType)
    +
    +  override def inputTypes: Seq[AbstractDataType] = Seq(NumericType, NumericType)
    +
    +  override def checkInputDataTypes(): TypeCheckResult =
    +    TypeUtils.checkForOrderingExpr(child.dataType, "function percentile")
    +
    +  override def supportsPartial: Boolean = false
    +
    +  override def aggBufferSchema: StructType = StructType.fromAttributes(aggBufferAttributes)
    +
    +  override val aggBufferAttributes: Seq[AttributeReference] = percentiles.map(percentile
=>
    +    AttributeReference(percentile.toString, DoubleType)())
    +
    +  override val inputAggBufferAttributes: Seq[AttributeReference] =
    +    aggBufferAttributes.map(_.newInstance())
    +
    +  override def initialize(buffer: MutableRow): Unit = {
    +    var i = 0
    +    while (i < percentiles.size) {
    +      buffer.setNullAt(mutableAggBufferOffset + i)
    +      i += 1
    +    }
    +  }
    +
    +  override def update(buffer: MutableRow, input: InternalRow): Unit = {
    +    val v = child.eval(input)
    +
    +    val key = v match {
    +      case o: Byte => o.toDouble
    +      case o: Short => o.toDouble
    +      case o: Int => o.toDouble
    +      case o: Long => o.toDouble
    +      case o: Float => o.toDouble
    +      case o: Decimal => o.toDouble
    +      case o: Double => o
    +      case _ => sys.error("Percentile is restricted to Numeric types only.")
    --- End diff --
    
    I don't this this is possible. Spark will try to provide numeric types (and inject casts
if needed), because you have defined `override def inputTypes: Seq[AbstractDataType] = Seq(NumericType,
NumericType)`


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org


Mime
View raw message