NGA-TRAN opened a new issue, #18258: URL: https://github.com/apache/datafusion/issues/18258
This task is part of feature #18249 Plan pruning can be adaptable, with several strategies depending on system constraints and goals. Here are a few examples: - **No pruning until the end:** If enumeration produces a manageable number of partial plans and memory is abundant, we can defer pruning entirely. A cost model is then applied at the final stage to select the best plans. - **Incremental pruning by level:** To prevent plan space explosion, we retain a subset of plans at each level, using a cost model to discard less promising ones. - **Memory-aware pruning:** Pruning is triggered only when memory usage exceeds a predefined threshold. - **Plan count-based pruning:** We prune only when the number of partial plans surpasses a specified limit. - **Hybrid pruning strategy:** only prune when certain conditions hit and prune to keep certain plans This flexibility allows pruning strategies to be tailored to system capacity and optimization goals. -- 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]
