JFinis commented on PR #221: URL: https://github.com/apache/parquet-format/pull/221#issuecomment-2938761824
I agree @orlp, all your arguments are valid and it's good to have the discussion. However, they all basically already came up in the [initial discussion](https://github.com/apache/parquet-format/pull/196), so they are merely another educated opinion rather than new points we didn't consider before and therefore aren't likely to sway existing opinions on the topic; or maybe they do. As mentioned, I'm happy with both approaches and willing to continue with any of the proposals. I mostly challenged your proposal to give you and others an overview of where the discussion went so far. While I like the conceptual simplicity of total ordering, I'm not a firm believer that it is the only possible solution to the problem. And I do recognize that your proposal of sorting NaN differently for min and max wasn't in my initial proposal and is probably a good idea, so if we went with NaN-counts, there are things that could be improved upon in the initial proposal before continuing. So from my side, I'm happy to continue either way. But I guess we now should at least get an informal vote going to see where the majority opinion resides on this. -- 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]
