[ https://issues.apache.org/jira/browse/CALCITE-5717?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Julian Hyde resolved CALCITE-5717. ---------------------------------- Resolution: Fixed Fixed in [379f41d3|https://github.com/apache/calcite/commit/379f41d3be465992328d5659ea62b8355e0399d1]. That fix performs the rewrite in {{RelBuilder.project}}; we could still accept similar functionality via {{ProjectAggregateMergeRule}} if it is useful. > Prune Project's input if project has no InputRef and the input has 1 row > ------------------------------------------------------------------------ > > Key: CALCITE-5717 > URL: https://issues.apache.org/jira/browse/CALCITE-5717 > Project: Calcite > Issue Type: Improvement > Components: core > Reporter: Benchao Li > Assignee: Benchao Li > Priority: Major > Labels: pull-request-available > Fix For: 1.35.0 > > > Prune {{Project}}'s input if project has no {{RexInputRef}} and the input has > 1 row. The idea is to improve SQL > {code:sql} > select 1 from emps group by () > {code} > In this case, we can safely prune the Aggregate. To be more specifically, I'm > proposing to match a Project -> Aggregate, and the Project has not field > reference from the Aggregate, and the Aggregate's grouping list is empty > (which means it will produce 1 row exactly), in this case, we can optimize it > to Project -> Values(1) to eliminate the Aggregate. > This idea is brought up in the discussion of CALCITE-5506 > We can add this improvement to {{ProjectAggregateMergeRule}}. -- This message was sent by Atlassian Jira (v8.20.10#820010)