[lake/paimon] Skip readable snapshot recomputation when latest compacted snapshot already in ZK

[luoyuxia]

Purpose

Linked issue: close #2626

Optimize readable snapshot computation performance: skip redundant tiered and readable offset recalculation when the latest compacted snapshot is already registered in Fluss (ZK). This avoids expensive recomputation when multiple APPEND snapshots follow a single COMPACT snapshot.

Brief change log

• Add ZK lookup logic in DvTableReadableSnapshotRetriever.computeReadableSnapshotAndOffsets()
• Return null directly if compacted snapshot already exists in ZK (no update needed)

Tests

• Add test case to verify that getReadableSnapshotAndOffsets() correctly returns null when compacted snapshot is already registered in ZK
• Existing tests continue to verify the normal flow

N/A

API and Format

N/A

Documentation

N/A

[Copilot]

Pull request overview

This PR aims to reduce redundant work in the Paimon DV readable-snapshot resolution by skipping recomputation when the latest compacted snapshot is already registered in Fluss (ZK), and adds a unit test to cover the new behavior.

Changes:

• Add a Fluss (ZK) lookup for the latest compacted snapshot and early-return when it is already present.
• Update method JavaDoc to document the new short-circuit behavior.
• Extend the existing readable-snapshot retriever test to assert the “skip recomputation” case.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
fluss-lake/fluss-lake-paimon/src/main/java/org/apache/fluss/lake/paimon/utils/DvTableReadableSnapshotRetriever.java	Adds ZK/Fluss snapshot existence check and returns null to skip recomputation when present.
fluss-lake/fluss-lake-paimon/src/test/java/org/apache/fluss/lake/paimon/utils/DvTableReadableSnapshotRetrieverTest.java	Adds a test scenario expecting null when the latest compacted snapshot is already registered.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[luoyuxia]

@wuchong I submit a pr to skip readable snapshot recomputation when the latest compacted snapshot is already registered in Fluss (ZK) for paimon deletion vector table. Could you please help review?