mihailom-db commented on code in PR #45383: URL: https://github.com/apache/spark/pull/45383#discussion_r1546046606
########## sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/CollationTypeCasts.scala: ########## @@ -0,0 +1,164 @@ +/* + * 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.analysis + +import javax.annotation.Nullable + +import scala.annotation.tailrec + +import org.apache.spark.sql.catalyst.expressions.{BinaryExpression, Cast, Collate, ComplexTypeMergingExpression, CreateArray, ExpectsInputTypes, Expression, Predicate, SortOrder} +import org.apache.spark.sql.errors.QueryCompilationErrors +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types.{AbstractDataType, ArrayType, DataType, StringType} + +object CollationTypeCasts extends TypeCoercionRule { + override val transform: PartialFunction[Expression, Expression] = { + case e if !e.childrenResolved => e + case sc @ (_: BinaryExpression Review Comment: Also, I do not believe incorporating string collations into core functions is the best way to go. If you take a look at findWiderCommonForTwo, it is implemented differently for AnsiTypeCoercion and TypeCoercion. Firstly both of those take leftFold and go through expressions in order, but in our case that is not what we want. We want to first know what collations are explicit and then look for anything else. This is important because we want to have clear design on which error has to come first, and explicit mismatch should always come before implicit one. Another thing is that AnsiTypeCoercion seems to behave differently for StringTypes, we do not reorder any type in the sequence, which would result in implementing the same thing we did here with addition of another rule, but just in a core, already well defined function. One more thing I would add is that not all rules actually use core functions, e.g. ConcatCoercion. Also doing collations casting in one rule is way more efficient, as otherwise we would have to reorder StringTypes in every other rule, to make sure explicit collation mismatches are thrown first, which would result in multiple reorderings as opposed to constant of 2, one at the beginning for all StringTypes and the other at the end for all expressions that were cast to StringType from other types in some other implicit casting rules. -- 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. To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org 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