Hello Flink community, Currently we are working FlinkSQL for real time processing in our team. While developing it, we understood that if the SQL is updated and want to rollback to previous savepoint if any error, it is not possible and throws error. While reading more about it, understood we have to assign operator ID for savepoint, but not sure how to set it through FlinkSQL. So wanted to take your support in resolving it, if any suggestions/ideas please let us know. Thank you. Notes: Found few references on internet, but couldn’t find a solution. 1. Adding operator could be done in Java. https://www.tabnine.com/code/java/classes/org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator 2. Not possible for SQL queries.[ from 2019 ] https://stackoverflow.com/questions/55464955/how-to-add-uid-to-operator-in-flink-table-api 3. “we don’t support stateful upgrades between Flink versions for Flink SQL yet” https://issues.apache.org/jira/browse/FLINK-28861

You might interested in the concept of Compiled Plan. And if you're looking for a way to extract and modify uids for SQL jobs, you can take a look at

org.apache.flink.table.planner.plan.nodes.exec.TransformationsTest#checkUids

Thank you, I will check them.