Enable `LocalTableScan` conversion by default

未关闭
#7,426 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

维护者通常 2 天内回复

还没有人认领这个 Issue。

评估

难度
4/5
预计耗时
3-5 天
新手友好度
25/100
Issue 类型
功能
描述清晰度
基本清楚
活跃度
活跃
技术栈
rust, scala, spark

调研方向

The flag spark.comet.exec.localTableScan.enabled gates the conversion of LocalTableScanExec, so start with the Scala rule that handles that operator and the config definition next to it. Read the linked failing Spark SQL tests from #4347 and #6795 to see what breaks when the flag is on. Done means the flag defaults to true, those tests pass, and a small-relation benchmark shows no slowdown.

由索引模型根据 Issue 内容生成。

描述

enhancement requires-triage
What is the problem the feature request solves?

LocalTableScanExec is disabled by default (spark.comet.exec.localTableScan.enabled). Small local relations, such as INSERT ... VALUES, createDataFrame(Seq(...)) and test tables, keep the operators above them in Spark. That includes native Iceberg and Parquet writes of that data.

Describe the potential solution

Check that converting LocalTableScanExec doesn't slow down common small-relation plans. Fix the Spark SQL test failures seen with it enabled (#4347, #6795), then enable it by default.

Additional context

The scan itself rarely gets faster. The point is to let the operators above it, especially native writes, convert.

主要语言
Scala
星标
1.3k
派生
414
平均合并
3 天 5 小时
30 天内合并 PR
210

环境准备

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

apache/datafusion-comet 的其他 Issue

查看 apache/datafusion-comet 的全部 Issue

相似的 Issue

更多 Scala Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。