lw309637554 commented on pull request #2136: URL: https://github.com/apache/hudi/pull/2136#issuecomment-741794311
> I thought of one more scenario. @lw309637554 Love to get your thoughts. Let's say the table was created with `GLOBAL_BLOOM` and then we allow writes with `BLOOM` (which can create duplication across partitions. then we also allow later writes with `GLOBAL_BLOOM`. The current checks bring a lot of value. Just pointing out that its not fully bulletproof. wdyt > > other than this, if we can think about metaClient creation, PR is ready @vinothchandar thanks. 1. about the GLOBAL_BLOOM-> BLOOM -> GLOBAL_BLOOM scenario have two option option 1: check GLOBAL_BLOOM incompatible with BLOOM strict. It will be concise option 2: update the index in hoodie properties when the index changed. it will be too heavy wdyt？ 2. about metaclient creation have resolved ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
