blaginin commented on PR #13177: URL: https://github.com/apache/datafusion/pull/13177#issuecomment-2453061327
And one more question on a slightly related topic. Regardless of the option we choose here, we’ll still have the same "triangles" on the profiling charts, where we don’t optimize the next item until we finish optimizing the previous one. <img width="1208" alt="Arc 2024-11-02 17 25 32" src="https://github.com/user-attachments/assets/90494bba-73f2-467b-a0aa-49b4cd22ee65";> Can the actual issue be not the stack size but the execution order in general? Surely _some_ optimisation rules could be applied to children in parallel, potentially giving a significant performance boost when optimising the query. Sorry if you’ve already discussed this anywhere, but I couldn’t find anything on GitHub. I get that not all optimisation rules can run in parallel, but if this challenge was taken into account, it feels like the whole context of this discussion would look pretty different 🤔 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
