pt:int id:int score:int tag:string
primary key: pt,id
partition key: pt
bucket count: 2 (both buckets populated: [0, 1])

Generated by Apache Paimon Java release-2.0.0.
file format: orc
deletion-vectors.enabled: true
deletion-vectors.merge-on-read: false
write-only: true (explicit fixture commits produce snapshots 2, 4, 5)
primary-key btree index: score

Snapshot semantics:
snapshot-1 APPEND: pt=1 and pt=2, ids 1..100 in each; score=(pt*100+id)%10; even tag=keep, odd tag=drop; 200 rows.
snapshot-2 COMPACT: full compaction of every populated partition/bucket builds source-backed score BTree payloads; 200 rows.
tag fallback-base and branch fallback are created from snapshot-2; the branch remains at its 200 indexed rows; it is used to validate fallback-branch TableRead routing.
snapshot-3 APPEND: add (id=101,score=0,tag=keep) to both partitions; indexed compacted files and visible unindexed APPEND-source files coexist; 202 rows.
snapshot-4 COMPACT: delete (pt=1,id=10), update (pt=2,id=10) to score=88/tag=updated, then lookup compact in one fixture commit; deletion vectors are present; 201 rows.
snapshot-5 COMPACT: full compaction rebuilds all partition/bucket BTree groups; 201 rows.

Fixture construction note: with deletion vectors enabled and merge-on-read disabled, Java and C++ batch scans skip literal level-0 files. Snapshot 3 therefore promotes Java-prepared files to the maximum manifest level while retaining FileSource.APPEND. The source-backed PK BTree policy indexes only COMPACT files above level 0, making these files visible but deliberately unindexed. Snapshot 4 publishes the Java-prepared compact outputs and index changes while omitting their transient, never-active append inputs. These are test-fixture constructions, not ordinary end-user write workflows.
