Github user cloud-fan commented on the issue: https://github.com/apache/spark/pull/21123 @rdblue I really appreciate your work for adding these new logical plans, and they are indeed incremental: when you add `AppendData`, you change `DataFrameWriter` to only create `AppendData` if `SaveMode` is "append". For other modes, still use the old `WriteToDataSourceV2`. That said, every PR we merged for data source v2, makes data source v2 better and still usable. I don't want to change this policy in #22009. I understand your concern of keeping the bad `SaveMode` API in data source v2, I hate it too. We should definitely revisit it before marking data source v2 as stable, but I don't think we need to rush to a decision in #22009 , which doesn't mark the v2 API stable. What do you think?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org