dtenedor commented on code in PR #36527: URL: https://github.com/apache/spark/pull/36527#discussion_r872694267
########## sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/RewriteNonAggregateFirst.scala: ########## @@ -0,0 +1,72 @@ +/* + * 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.{Alias, Literal} +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, First} +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.FIRST +import org.apache.spark.sql.types.IntegerType + +/** + * Rewrite aggregate plan with a single [[First]] function when grouping is absent. In such a case + * the query is equivalent to simple projection with limit 1. + * + * Input Pseudo-Query: + * {{{ + * SELECT FIRST(col) FROM table + * }}} + * + * Rewritten Query: + * {{{ + * SELECT col FROM table LIMIT 1 Review Comment: Good question, yes, this optimization only covers cases of one or more `FIRST` aggregate functions and only when `IGNORE NULLS` is not present, and no other aggregate functions are present. For other cases of the `FIRST` aggregate function, we plan to update the query execution to stop consuming input values after receiving the first one when possible (the current implementation consumes all the remaining values and discards them). The advantage of generating a logical `LIMIT 1` in this case is that the optimizer can recursively push it down. This can result in more efficient execution since e.g. data source scans can prune partitions or do other improvements outside of the aggregation operator itself. Re: complexity tradeoff vs. performance improvement, please leave opinions if needed. (Note that pruning on `_.containsPattern(FIRST)` will skip the rule iteration unless the query plan contains the FIRST aggregate function.) -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
