matthewmturner commented on issue #9964: URL: https://github.com/apache/arrow-datafusion/issues/9964#issuecomment-2063749535
Thats interesting that dashmap performed better. Would you mind also doing a query with a filter / pruning involved and comparing the results (perhaps with a range scan as well)? And just to confirm, the cache is filled at creation time during schema inference, right? Also is the cache used by both `ListingTable` and `ParquetExec`? Im not sure how it works if adding the cache removes the duplicate calls or if additional work is required for that (if it's additional work could probably be done as a separate PR once cache implementation is finalized). I think it would also be nice to have a new benchmark for this to help monitor moving forward. -- 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]
