feathr-ai/feathr

Allow `OUTPUT_PARALLELISM` and `OVERWRITE_MODE` in join output

Open

#91 建立於 2022年3月24日

在 GitHub 查看
 (2 留言) (0 反應) (0 負責人)Scala (1,929 star) (244 fork)batch import
buggood first issuehelp wanted

描述

Seems it's not honored per this line: https://github.com/linkedin/feathr/blob/main/src/main/scala/com/linkedin/feathr/offline/generation/SparkIOUtils.scala#L21

貢獻者指南

技術棧
scala
領域
data
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
needs investigation
前置要求
Basic ScalaFamiliarity with Spark output modes
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
25
研究方向
Start by examining the referenced line (SparkIOUtils.scala:21) and trace how OUTPUT PARALLELISM and OVERWRITE MODE are currently handled in the output path. Check if similar settings are applied in other parts of the Feathr codebase, such as for non join outputs. Review any existing tests or comments in the issue to understand the expected behavior. If possible, look at recent commits or PRs that might have addressed this partially.
Allow `OUTPUT_PARALLELISM` and `OVERWRITE_MODE` in join output · feathr-ai/feathr#91 | Good First Issue