geserdugarov commented on PR #13216: URL: https://github.com/apache/hudi/pull/13216#issuecomment-3026908033
> @geserdugarov what is the performance concern that it may induces in this PR? We should definitely fix it IMO, can you guide me the lines in this patch? > > In general, all the changes in this path should not bring in performance issues, some hook/empty invocations shoud be fine and the cost could be negligible. > > I agree we should limit the code complexity, it's the way we are evolving the project, we welcome new features always if it is valuable, then try to make the code in good shape as much as we can, but yeah, sometimes it is hard to give a elegant solution based on the amout of work it could take. > > Appreciate your points. @danny0405 , from quick checking of this patch, you're right, looks like there should be no performance issues. I only see that we started to pass `HoodieTableMetaClient` way more, and it's a `Serializable` class, which contains a lot of not `transient` objects. If there are any frequently called place in the code, where we will start to use Kryo SerDe of this class, (which I don't actually see) it could potentially decrease performance. But again, it would be great to have some continuous performance monitoring. I'll try to think about it in my spare time. -- 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]
