Support Iceberg's `MergeRowsExec` on Spark 3.4 and 3.5

未关闭
#7,427 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
1-2 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

Start from the existing Spark 4.0 serde for MergeRowsExec in the Comet Scala sources, then find where Iceberg's Spark 3.4 and 3.5 MergeRowsExec from iceberg-spark-extensions would need a version-specific shim. Read #5122, #5644 and #6240 for how row-level plans fit together. Done means the operator converts natively on Spark 3.4 and 3.5 with matched, not-matched and cardinality-check behaviour preserved.

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?

Comet runs MergeRowsExec natively on Spark 4.0 and later. On Spark 3.4 and 3.5, MERGE INTO on Iceberg tables plans Iceberg's own MergeRowsExec from iceberg-spark-extensions, which Comet does not convert. The operators above it, including the copy-on-write ReplaceData write, then stay in Spark.

Describe the potential solution

Add a serde and shim for Iceberg's Spark 3.4 and 3.5 MergeRowsExec that reuses the native merge-rows operator from the Spark 4.0 path. Cover matched and not-matched clauses, conditions on each, and the cardinality check.

Additional context

Related: #5122 (row-level MERGE / UPDATE / DELETE plans). This only pays off fully together with native ReplaceData (#5644) and WriteDelta (#6240).

主要语言
Scala
星标
1.3k
派生
414
平均合并
3 天 4 小时
30 天内合并 PR
202

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。