carbondata-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ravipesala <...@git.apache.org>
Subject [GitHub] incubator-carbondata pull request #805: [CARBONDATA-934] Cast Filter Express...
Date Fri, 21 Apr 2017 10:29:39 GMT
Github user ravipesala commented on a diff in the pull request:

    https://github.com/apache/incubator-carbondata/pull/805#discussion_r112659542
  
    --- Diff: integration/spark2/src/main/scala/org/apache/spark/sql/optimizer/CarbonFilters.scala
---
    @@ -0,0 +1,380 @@
    +/*
    + * 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.carbondata.spark
    +
    +import java.text.SimpleDateFormat
    +import java.util.Date
    +
    +import scala.collection.mutable.ArrayBuffer
    +
    +import org.apache.spark.sql.catalyst.expressions._
    +import org.apache.spark.sql.catalyst.expressions.aggregate.AggregateExpression
    +import org.apache.spark.sql.execution.CastExpressionOptimization
    +import org.apache.spark.sql.optimizer.AttributeReferenceWrapper
    +import org.apache.spark.sql.sources
    +import org.apache.spark.sql.sources.Filter
    +import org.apache.spark.sql.types._
    +
    +import org.apache.carbondata.core.constants.CarbonCommonConstants
    +import org.apache.carbondata.core.keygenerator.directdictionary.timestamp.TimeStampDirectDictionaryGenerator
    +import org.apache.carbondata.core.metadata.datatype.DataType
    +import org.apache.carbondata.core.metadata.schema.table.CarbonTable
    +import org.apache.carbondata.core.metadata.schema.table.column.CarbonColumn
    +import org.apache.carbondata.core.scan.expression.{ColumnExpression => CarbonColumnExpression,
Expression => CarbonExpression, LiteralExpression => CarbonLiteralExpression}
    +import org.apache.carbondata.core.scan.expression.conditional._
    +import org.apache.carbondata.core.scan.expression.logical.{AndExpression, FalseExpression,
OrExpression}
    +import org.apache.carbondata.spark.util.CarbonScalaUtil
    +
    +
    +/**
    + * All filter conversions are done here.
    + */
    +object CarbonFilters {
    +
    +
    +  /**
    +   * Converts data sources filters to carbon filter predicates.
    +   */
    +  def createCarbonFilter(schema: StructType,
    +      predicate: sources.Filter): Option[CarbonExpression] = {
    +    val dataTypeOf = schema.map(f => f.name -> f.dataType).toMap
    +
    +    def createFilter(predicate: sources.Filter): Option[CarbonExpression] = {
    +      predicate match {
    +
    +        case sources.EqualTo(name, value) =>
    +          Some(new EqualToExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.Not(sources.EqualTo(name, value)) =>
    +          Some(new NotEqualsExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.EqualNullSafe(name, value) =>
    +          Some(new EqualToExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.Not(sources.EqualNullSafe(name, value)) =>
    +          Some(new NotEqualsExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.GreaterThan(name, value) =>
    +          Some(new GreaterThanExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.LessThan(name, value) =>
    +          Some(new LessThanExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.GreaterThanOrEqual(name, value) =>
    +          Some(new GreaterThanEqualToExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.LessThanOrEqual(name, value) =>
    +          Some(new LessThanEqualToExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, value)))
    +        case sources.In(name, values) =>
    +          Some(new InExpression(getCarbonExpression(name),
    +            new ListExpression(
    +              convertToJavaList(values.map(f => getCarbonLiteralExpression(name, f)).toList))))
    +        case sources.Not(sources.In(name, values)) =>
    +          Some(new NotInExpression(getCarbonExpression(name),
    +            new ListExpression(
    +              convertToJavaList(values.map(f => getCarbonLiteralExpression(name, f)).toList))))
    +        case sources.IsNull(name) =>
    +          Some(new EqualToExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, null), true))
    +        case sources.IsNotNull(name) =>
    +          Some(new NotEqualsExpression(getCarbonExpression(name),
    +            getCarbonLiteralExpression(name, null), true))
    +        case sources.And(lhs, rhs) =>
    +          (createFilter(lhs) ++ createFilter(rhs)).reduceOption(new AndExpression(_,
_))
    +        case sources.Or(lhs, rhs) =>
    +          for {
    +            lhsFilter <- createFilter(lhs)
    +            rhsFilter <- createFilter(rhs)
    +          } yield {
    +            new OrExpression(lhsFilter, rhsFilter)
    +          }
    +        case CastExpr(expr: Expression) =>
    +          Some(transformExpression(expr))
    +        case _ => None
    +      }
    +    }
    +
    +    def getCarbonExpression(name: String) = {
    +      new CarbonColumnExpression(name,
    +        CarbonScalaUtil.convertSparkToCarbonDataType(dataTypeOf(name)))
    +    }
    +
    +    def getCarbonLiteralExpression(name: String, value: Any): CarbonExpression = {
    +      val dataTypeOfAttribute = CarbonScalaUtil.convertSparkToCarbonDataType(dataTypeOf(name))
    +      val dataType = if (Option(value).isDefined
    +                         && dataTypeOfAttribute == DataType.STRING
    +                         && value.isInstanceOf[Double]) {
    +        DataType.DOUBLE
    +      } else {
    +        dataTypeOfAttribute
    +      }
    +      new CarbonLiteralExpression(value, dataType)
    +    }
    +
    +    createFilter(predicate)
    +  }
    +
    +
    +  // Check out which filters can be pushed down to carbon, remaining can be handled in
spark layer.
    +  // Mostly dimension filters are only pushed down since it is faster in carbon.
    +  def selectFilters(filters: Seq[Expression],
    +      attrList: java.util.HashSet[AttributeReferenceWrapper],
    +      aliasMap: CarbonAliasDecoderRelation): Unit = {
    +    def translate(expr: Expression, or: Boolean = false): Option[sources.Filter] = {
    +      expr match {
    +        case or@Or(left, right) =>
    +
    +          val leftFilter = translate(left, or = true)
    +          val rightFilter = translate(right, or = true)
    +          if (leftFilter.isDefined && rightFilter.isDefined) {
    +            Some(sources.Or(leftFilter.get, rightFilter.get))
    +          } else {
    +            or.collect {
    +              case attr: AttributeReference =>
    +                attrList.add(AttributeReferenceWrapper(aliasMap.getOrElse(attr, attr)))
    +            }
    +            None
    +          }
    +        case And(left, right) =>
    +          (translate(left) ++ translate(right)).reduceOption(sources.And)
    +        case EqualTo(a: Attribute, Literal(v, t)) =>
    +          Some(sources.EqualTo(a.name, v))
    +        case EqualTo(l@Literal(v, t), a: Attribute) =>
    +          Some(sources.EqualTo(a.name, v))
    +        case c@EqualTo(Cast(a: Attribute, _), Literal(v, t)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@EqualTo(Literal(v, t), Cast(a: Attribute, _)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case Not(EqualTo(a: Attribute, Literal(v, t))) =>
    +          Some(sources.Not(sources.EqualTo(a.name, v)))
    +        case Not(EqualTo(Literal(v, t), a: Attribute)) =>
    +          Some(sources.Not(sources.EqualTo(a.name, v)))
    +        case c@Not(EqualTo(Cast(a: Attribute, _), Literal(v, t))) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@Not(EqualTo(Literal(v, t), Cast(a: Attribute, _))) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case IsNotNull(a: Attribute) =>
    +          Some(sources.IsNotNull(a.name))
    +        case IsNull(a: Attribute) =>
    +          Some(sources.IsNull(a.name))
    +        case Not(In(a: Attribute, list)) if !list.exists(!_.isInstanceOf[Literal]) =>
    +          val hSet = list.map(e => e.eval(EmptyRow))
    +          Some(sources.Not(sources.In(a.name, hSet.toArray)))
    +        case In(a: Attribute, list) if !list.exists(!_.isInstanceOf[Literal]) =>
    +          val hSet = list.map(e => e.eval(EmptyRow))
    +          Some(sources.In(a.name, hSet.toArray))
    +        case c@Not(In(Cast(a: Attribute, _), list)) if !list.exists(!_.isInstanceOf[Literal])
=>
    +          Some(CastExpr(c))
    +        case c@In(Cast(a: Attribute, _), list) if !list.exists(!_.isInstanceOf[Literal])
=>
    +            Some(CastExpr(c))
    +        case GreaterThan(a: Attribute, Literal(v, t)) =>
    +          Some(sources.GreaterThan(a.name, v))
    +        case GreaterThan(Literal(v, t), a: Attribute) =>
    +          Some(sources.LessThan(a.name, v))
    +        case c@GreaterThan(Cast(a: Attribute, _), Literal(v, t)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@GreaterThan(Literal(v, t), Cast(a: Attribute, _)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case LessThan(a: Attribute, Literal(v, t)) =>
    +          Some(sources.LessThan(a.name, v))
    +        case LessThan(Literal(v, t), a: Attribute) =>
    +          Some(sources.GreaterThan(a.name, v))
    +        case c@LessThan(Cast(a: Attribute, _), Literal(v, t)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@LessThan(Literal(v, t), Cast(a: Attribute, _)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case GreaterThanOrEqual(a: Attribute, Literal(v, t)) =>
    +          Some(sources.GreaterThanOrEqual(a.name, v))
    +        case GreaterThanOrEqual(Literal(v, t), a: Attribute) =>
    +          Some(sources.LessThanOrEqual(a.name, v))
    +        case c@GreaterThanOrEqual(Cast(a: Attribute, _), Literal(v, t)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@GreaterThanOrEqual(Literal(v, t), Cast(a: Attribute, _)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case LessThanOrEqual(a: Attribute, Literal(v, t)) =>
    +          Some(sources.LessThanOrEqual(a.name, v))
    +        case LessThanOrEqual(Literal(v, t), a: Attribute) =>
    +          Some(sources.GreaterThanOrEqual(a.name, v))
    +        case c@LessThanOrEqual(Cast(a: Attribute, _), Literal(v, t)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@LessThanOrEqual(Literal(v, t), Cast(a: Attribute, _)) =>
    +          CastExpressionOptimization.checkIfCastCanBeRemove(c)
    +        case c@Cast(a: Attribute, _) =>
    +          Some(CastExpr(c))
    +        case others =>
    +          if (!or) {
    +            others.collect {
    +              case attr: AttributeReference =>
    +                attrList.add(AttributeReferenceWrapper(aliasMap.getOrElse(attr, attr)))
    +            }
    +          }
    +          None
    +      }
    +    }
    +
    +    filters.flatMap(translate(_, false)).toArray
    +  }
    +
    +  def transformExpression(expr: Expression): CarbonExpression = {
    +    expr match {
    +      case Or(left, right)
    +        if (isCarbonSupportedDataTypes(left) && isCarbonSupportedDataTypes(right))
=> new
    --- End diff --
    
    remove unnecessary extra parenthesis in if conditions for all cases.


---
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.
---

Mime
View raw message