steveloughran commented on PR #15153: URL: https://github.com/apache/iceberg/pull/15153#issuecomment-3807449592
Removing 3.x support would be best -but what's a realistic timeline for that? I really don't want to get into the situation of having to permanently rebase the code as that's always a nightmare -too much risk of incompatible changes in HadoopFileIO which I'd be left to deal with. The Trash stuff has already done that, which I'm doing a second iteration on in #15111 so that it'll be stable from then on. I also have an end-to-end/regression test in https://github.com/apache/hadoop/pull/7316 which depends on a public Iceberg release with the feature; it's got the same problem of staying up to date. If 1.11 is going to cut 3.4 then all is good; otherwise that downgrade should go in for safety and I'll get back to the reflection-based approach to the bulk delete with a goal of it in 1.11 and working only on runtimes with the new feature. FWIW we have done that in parquet for vector IO/ scatter/gather reads, as we like that 30% speedup on queries. -- 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]
