[Feature] Roadmap for Paimon C++ 0.5.0 #519

Search before asking

• I searched in the issues and found nothing similar.

Motivation

Paimon C++ 0.5.0 will focus on completing Data Evolution maintenance, expanding global and file index capabilities, making all four real-time modes usable, and closing important compatibility gaps with Paimon Java.

Solution

1. Data Evolution

• Support compaction across evolved column groups, including tables with deletion vectors.
• Support dedicated BLOB and VECTOR file compaction and improve file lifecycle management.

2. Global Index

• Complete global index support for primary-key tables.
• Support additional index types, e.g., multivalue and FM indexes.
• Improve index maintenance, coverage handling, and cleanup.

3. Real-Time Support

• Complete append, PK DV, PK MOR, and append deduplicate modes.
• Support in-memory scalar, vector, and full-text indexes.
• Improve consistency, recovery, and memory reclamation.

4. File Index

• Support hybrid retrieval combining scalar filters, vector search, and full-text search.
• Complete predicate execution, score output, and cross-file TopK merging.
• Add tasks for building, backfilling, and rebuilding file indexes.

5. Core Feature Alignment

• Table API: Introduce a public Table abstraction for creating Read/Write/Scan/Commit operators, and consolidate operator creation logic in the Table layer.
• Performance Optimization: Decouple prefetch from warmup.
• Complete TIME type support.
• Support dynamic-bucket primary-key tables.
• Support cross-partition upserts.

Anything else?

Follow-up to #186. Each item should include focused tests, documentation, and Java/C++ compatibility validation where applicable.

Are you willing to submit a PR?

• I'm willing to submit a PR!