slfan1989 commented on issue #12762: URL: https://github.com/apache/iceberg/issues/12762#issuecomment-2795599098
@pvary Although Hive provides a good solution, after careful consideration, I still hope to seize an opportunity to implement a similar functionality in Iceberg’s Spark Procedure, offering more choices to users. From a user’s perspective, our online environment uses Hive version 2.3.9, and upgrading to a new version is somewhat challenging due to the need for extensive compatibility evaluations. Additionally, there are significant code differences between Hive 2.3.9 and Hive 4.0.0. Therefore, I can only try to backport the relevant PRs to our internal branch following the approach of HIVE-25008. Since Hive’s HMS is a stateful service, the upgrade process will take some time. Considering that 90% of our SQL tasks run on Spark, and the changes to Iceberg will eventually be compiled into a Spark runtime package, we can quickly update the Spark client to implement this functionality. After a comprehensive evaluation, using Spark’s stored procedure seems to be the more straightforward option. To facilitate discussion, I will submit a draft PR that outlines the implementation approach. The changes are relatively minor and primarily reuse existing Iceberg capabilities. cc: @nastra @szehon-ho -- 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]
