aokolnychyi edited a comment on pull request #2048: URL: https://github.com/apache/iceberg/pull/2048#issuecomment-758714296
@rdblue @shardulm94 @RussellSpitzer, not related to this PR but a general question: the metrics config relies on column names instead of column ids as it is controlled through table properties. I was wondering whether it is safe to do so. I considered the following cases: - While writing new data, it should be always OK to follow the name-based approach if the config is up-to-date. - While importing data (assuming we have a correct name mapping), we will assign correct ids to columns, use the ids to find the current aliases in the schema, so this should work. - When fixing (re-importing) metadata (assuming we have renamed a column with a reliable id), we will just use the column id in the file to find the current name, so this should work. Overall, it seems safe to me. Any cases I missed? ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
