I'm experimenting with a Nessie + trino + iceberg implementation.
But i'd like to bulk ingest data to a table using flink, while performing deletes on the same table using trino, and i'd like to run nessie gc every 24h or so and remove the files from storage, all eventually without manual input and without corrupting the table.
Is this at all possible?
I was thinking something along the lines of
- the ingestion job running on main
- the delete job creates a branch and merges on completion
- the gc job runs routinely, expires snapshots and removes orphan files
am i barking up the wrong tree?