Perform Iceberg table branching and tagging to manage release workflows and long-running experiments

Verified steps

1. Create a branch for an experimental write using CALL catalog.system.create_branch(table => 'db.table', name => 'experiment-1', snapshot_id => <current_snapshot_id>) or via the Iceberg Java API with table.manageSnapshots().createBranch('experiment-1').commit()
2. Write data to the branch by setting the write option spark.wap.branch to the branch name in SparkSQL or passing branch in write options; reads from other branches are isolated
3. Validate data quality on the branch with queries scoped to it using the branch time-travel syntax: SELECT * FROM db.table VERSION AS OF 'experiment-1'
4. Promote the branch to main by fast-forwarding: CALL catalog.system.fast_forward(table => 'db.table', branch => 'main', to => 'experiment-1') — only works if main has not advanced past the branch's base
5. Create a tag to pin a known-good snapshot for audit or rollback: CALL catalog.system.create_tag(table => 'db.table', name => 'v2024-01-release', snapshot_id => <snapshot_id>, max_snapshot_age_ms => 0) so the snapshot is retained even after expire_snapshots runs
6. Drop stale branches and tags when no longer needed: CALL catalog.system.remove_branch / remove_tag; verify retention policy on the catalog side to avoid accumulating indefinitely retained snapshots