sunchao commented on a change in pull request #29565: URL: https://github.com/apache/spark/pull/29565#discussion_r488112189
########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/UnwrapCastInBinaryComparison.scala ########## @@ -0,0 +1,236 @@ +/* + * 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.optimizer + +import org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.Literal.FalseLiteral +import org.apache.spark.sql.catalyst.plans.logical.LogicalPlan +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.types._ + +/** + * Unwrap casts in binary comparison operations with patterns like following: + * + * `BinaryComparison(Cast(fromExp, toType), Literal(value, toType))` + * or + * `BinaryComparison(Literal(value, toType), Cast(fromExp, toType))` + * + * This rule optimizes expressions with the above pattern by either replacing the cast with simpler + * constructs, or moving the cast from the expression side to the literal side, which enables them + * to be optimized away later and pushed down to data sources. + * + * Currently this only handles cases where: + * 1). `fromType` (of `fromExp`) and `toType` are of integral types (i.e., byte, short, int and + * long) + * 2). `fromType` can be safely coerced to `toType` without precision loss (e.g., short to int, + * int to long, but not long to int) + * + * If the above conditions are satisfied, the rule checks to see if the literal `value` is within + * range `(min, max)`, where `min` and `max` are the minimum and maximum value of `fromType`, + * respectively. If this is true then it means we can safely cast `value` to `fromType` and thus + * able to move the cast to the literal side. That is: + * + * `cast(fromExp, toType) op value` ==> `fromExp op cast(value, fromType)` + * + * If the `value` is not within range `(min, max)`, the rule breaks the scenario into different + * cases and try to replace each with simpler constructs. + * + * if `value > max`, the cases are of following: + * - `cast(fromExp, toType) > value` ==> if(isnull(fromExp), null, false) + * - `cast(fromExp, toType) >= value` ==> if(isnull(fromExp), null, false) + * - `cast(fromExp, toType) === value` ==> if(isnull(fromExp), null, false) + * - `cast(fromExp, toType) <=> value` ==> false (if `fromExp` is deterministic) + * - `cast(fromExp, toType) <=> value` ==> cast(fromExp, toType) <=> value (if `fromExp` is + * non-deterministic) + * - `cast(fromExp, toType) <= value` ==> if(isnull(fromExp), null, true) + * - `cast(fromExp, toType) < value` ==> if(isnull(fromExp), null, true) + * + * if `value == max`, the cases are of following: + * - `cast(fromExp, toType) > value` ==> if(isnull(fromExp), null, false) + * - `cast(fromExp, toType) >= value` ==> fromExp == max + * - `cast(fromExp, toType) === value` ==> fromExp == max + * - `cast(fromExp, toType) <=> value` ==> fromExp <=> max + * - `cast(fromExp, toType) <= value` ==> if(isnull(fromExp), null, true) + * - `cast(fromExp, toType) < value` ==> fromExp =!= max + * + * Similarly for the cases when `value == min` and `value < min`. + * + * Further, the above `if(isnull(fromExp), null, false)` is represented using conjunction + * `and(isnull(fromExp), null)`, to enable further optimization and filter pushdown to data sources. + * Similarly, `if(isnull(fromExp), null, true)` is represented with `or(isnotnull(fromExp), null)`. + */ +object UnwrapCastInBinaryComparison extends Rule[LogicalPlan] { + override def apply(plan: LogicalPlan): LogicalPlan = plan transform { + case l: LogicalPlan => + l transformExpressionsUp { + case e @ BinaryComparison(_, _) => unwrapCast(e) + } + } + + private def unwrapCast(exp: Expression): Expression = exp match { + // Not a canonical form. In this case we first canonicalize the expression by swapping the + // literal and cast side, then process the result and swap the literal and cast again to + // restore the original order. + case BinaryComparison(Literal(_, literalType), Cast(fromExp, toType, _)) + if canImplicitlyCast(fromExp, toType, literalType) => + def swap(e: Expression): Expression = e match { + case GreaterThan(left, right) => LessThan(right, left) + case GreaterThanOrEqual(left, right) => LessThanOrEqual(right, left) + case EqualTo(left, right) => EqualTo(right, left) + case EqualNullSafe(left, right) => EqualNullSafe(right, left) + case LessThanOrEqual(left, right) => GreaterThanOrEqual(right, left) + case LessThan(left, right) => GreaterThan(right, left) + case _ => e + } + + swap(unwrapCast(swap(exp))) + + // In case both sides have integral type, optimize the comparison by removing casts or + // moving cast to the literal side. + case be @ BinaryComparison( + Cast(fromExp, toType: IntegralType, _), Literal(value, literalType)) + if canImplicitlyCast(fromExp, toType, literalType) => + simplifyIntegralComparison(be, fromExp, toType, value) + + case _ => exp + } + + /** + * Check if the input `value` is within range `(min, max)` of the `fromType`, where `min` and + * `max` are the minimum and maximum value of the `fromType`. If the above is true, this + * optimizes the expression by moving the cast to the literal side. Otherwise if result is not + * true, this replaces the input binary comparison `exp` with simpler expressions. + */ + private def simplifyIntegralComparison( + exp: BinaryComparison, + fromExp: Expression, + toType: IntegralType, + value: Any): Expression = { + + val fromType = fromExp.dataType + val (min, max) = getRange(fromType) + val (minInToType, maxInToType) = { + (Cast(Literal(min), toType).eval(), Cast(Literal(max), toType).eval()) + } + val ordering = toType.ordering.asInstanceOf[Ordering[Any]] + val minCmp = ordering.compare(value, minInToType) + val maxCmp = ordering.compare(value, maxInToType) + + if (maxCmp > 0) { + exp match { + case EqualTo(_, _) | GreaterThan(_, _) | GreaterThanOrEqual(_, _) => + falseIfNotNull(fromExp) + case LessThan(_, _) | LessThanOrEqual(_, _) => + trueIfNotNull(fromExp) + // make sure the expression is evaluated if it is non-deterministic + case EqualNullSafe(_, _) if exp.deterministic => + FalseLiteral + case _ => exp + } + } else if (maxCmp == 0) { + exp match { + case GreaterThan(_, _) => + falseIfNotNull(fromExp) + case LessThanOrEqual(_, _) => + trueIfNotNull(fromExp) + case LessThan(_, _) => + Not(EqualTo(fromExp, Literal(max, fromType))) + case GreaterThanOrEqual(_, _) | EqualTo(_, _) => + EqualTo(fromExp, Literal(max, fromType)) + case EqualNullSafe(_, _) => + EqualNullSafe(fromExp, Literal(max, fromType)) + case _ => exp + } + } else if (minCmp < 0) { + exp match { + case GreaterThan(_, _) | GreaterThanOrEqual(_, _) => + trueIfNotNull(fromExp) + case LessThan(_, _) | LessThanOrEqual(_, _) | EqualTo(_, _) => + falseIfNotNull(fromExp) + // make sure the expression is evaluated if it is non-deterministic + case EqualNullSafe(_, _) if exp.deterministic => + FalseLiteral + case _ => exp + } + } else if (minCmp == 0) { + exp match { + case LessThan(_, _) => + falseIfNotNull(fromExp) + case GreaterThanOrEqual(_, _) => + trueIfNotNull(fromExp) + case GreaterThan(_, _) => + Not(EqualTo(fromExp, Literal(min, fromType))) + case LessThanOrEqual(_, _) | EqualTo(_, _) => + EqualTo(fromExp, Literal(min, fromType)) + case EqualNullSafe(_, _) => + EqualNullSafe(fromExp, Literal(min, fromType)) + case _ => exp + } + } else { + // This means `value` is within range `(min, max)`. Optimize this by moving the cast to the + // literal side. + val lit = Cast(Literal(value), fromType) + exp match { + case GreaterThan(_, _) => GreaterThan(fromExp, lit) + case GreaterThanOrEqual(_, _) => GreaterThanOrEqual(fromExp, lit) + case EqualTo(_, _) => EqualTo(fromExp, lit) + case EqualNullSafe(_, _) => EqualNullSafe(fromExp, lit) + case LessThan(_, _) => LessThan(fromExp, lit) + case LessThanOrEqual(_, _) => LessThanOrEqual(fromExp, lit) + case _ => exp + } + } + } + + /** + * Check if the input `fromExp` can be safely cast to `toType` without any loss of precision, + * i.e., the conversion is injective. Note this only handles the case when both sides are of + * integral type. + */ + private def canImplicitlyCast(fromExp: Expression, toType: DataType, Review comment: Thanks. I'll check `foldable` in the follow-up PR to handle more types. And also fix the indentation. ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org