JFinis commented on PR #221: URL: https://github.com/apache/parquet-format/pull/221#issuecomment-3143774752
My personal opinion would be not to revive nan counts. Can we maybe first establish whether we have a majority for anything? I feel like we're being startled by a singular opinion here, while we already had broad consensus of switching from nan_counts to total order; that's why I created this second PR in the first place. It feels like going in circles without adding new insights. One technical comment: > I think the current order we have in TypeDefinedOrder is honestly fine: I don't think it's fully fine. It's not a total order, so it makes `sorting_columns` ambiguous (are NaNs first? Last? It's underdefined). Therefore, one cannot leverage sortedness on floating point columns. -- 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]
