[GitHub] [arrow] joosthooz commented on pull request #33738: GH-33737: [C++] simplify exec plan tracing

2023-01-23 Thread via GitHub
joosthooz commented on PR #33738: URL: https://github.com/apache/arrow/pull/33738#issuecomment-1400316462 Hi, I gave this branch a spin, and it seems that the nesting has become inconsistent:

[GitHub] [arrow] joosthooz commented on pull request #33738: GH-33737: [C++] simplify exec plan tracing

2023-01-19 Thread GitBox
joosthooz commented on PR #33738: URL: https://github.com/apache/arrow/pull/33738#issuecomment-1396819003 Nice, I think the WriteAndCheckBackpressure span is important because that's where the backpressure is checked and also it performs some work combining staged batches (in

[GitHub] [arrow] joosthooz commented on pull request #33738: GH-33737: [C++] simplify exec plan tracing

2023-01-18 Thread GitBox
joosthooz commented on PR #33738: URL: https://github.com/apache/arrow/pull/33738#issuecomment-1387311099 In the figure, why does `WaitForFinish(SinkNode:)` end earlier than the `ScalarAggregate`? Can we add a name (and maybe even an id number in case there are multiple) to the names so