Want to get some advises on schema evolution for Apache Paimon tables.
Currently we are onboarding Paimon in way:
- There is a Flink application which injects data to Paimon tables
- There are readers which consumes the data later
During experimenting with schema evolution noticed that when schema change was done in external way Flink application didn't pick up changed properties even after it was restarted. For application to pick up the change I had to restart job without its state. What riese a question how to properly manage the schema evolution, should that be imbedded to the application as pre job execution steps or should that be managed somehow else?
