refactor: unify BinarySection classes to single MemorySegment model and use string_view to avoid copies

[lxy-9602]

Purpose

Linked issue: #93

1. Refactors the internal row structures (BinarySection, BinaryString, BinaryRow, BinaryArray, BinaryMap) from a multi-segment model (std::vector<MemorySegment>) to a single-segment model (MemorySegment), ensuring contiguous memory layout. It also introduces GetStringView / string_view-based read paths to avoid unnecessary data copies in serialization and comparison operations.
2. CompactStrategy supports BucketedDvMaintainer.

Tests

CompactStrategyTest, TestPickFullCompaction
DataDefineTest, GetStringView
LookupMergeTreeCompactRewriterTest, TestRewriteWithDvAndAggForStringFields

API and Format

Blob::ArrowField() remove nullable param as blob not support null now.

Generative AI tooling

Partially Generated-by: Claude-4.6-Opus

[Copilot]

Pull request overview

Refactors serialization and compaction selection to reduce copies by using std::string_view in hot paths, and extends full-compaction picking logic to account for deletion vectors via BucketedDvMaintainer.

Changes:

• Enable use_view=true for InternalRow field getters and add string-view based write paths in RowCompactedSerializer / serializer utilities.
• Extend CompactStrategy::PickFullCompaction to accept a BucketedDvMaintainer and trigger rewrites when deletion vectors exist.
• Update/add unit tests to cover the new compaction selection behavior and adjust persist processor test setup.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 10 comments.

Show a summary per file

File	Description
src/paimon/core/mergetree/lookup/persist_processor_test.cpp	Reworks test fixture setup to build KeyValue from Arrow/ColumnarRow instead of BinaryRowGenerator.
src/paimon/core/mergetree/compact/compact_strategy_test.cpp	Updates PickFullCompaction call sites for new signature and adds DV-maintainer coverage.
src/paimon/core/mergetree/compact/compact_strategy.h	Adds DV maintainer parameter and logic to include files with deletion vectors in full compaction.
src/paimon/common/data/serializer/row_compacted_serializer.h	Adds RowWriter::WriteStringView to write length-prefixed bytes from std::string_view.
src/paimon/common/data/serializer/row_compacted_serializer.cpp	Enables view-based getters and removes intermediate allocations/copies when serializing string/binary.
src/paimon/common/data/serializer/binary_serializer_utils.cpp	Switches STRING/BINARY serialization to pass views instead of copying.
src/paimon/common/data/generic_row.h	Adds conversion path when a GenericRow field is stored as std::string_view but API requires BinaryString.

---

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

[Copilot]

Pull request overview

Copilot reviewed 66 out of 66 changed files in this pull request and generated 5 comments.

---

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

[Copilot]

Pull request overview

Copilot reviewed 66 out of 66 changed files in this pull request and generated 7 comments.

Comments suppressed due to low confidence (2)

src/paimon/core/mergetree/compact/compact_strategy.h:1

• Including bucketed_dv_maintainer.h in this header can significantly increase compile-time coupling. Since the API only needs std::shared_ptr<BucketedDvMaintainer>, you can forward-declare class BucketedDvMaintainer; in the paimon namespace and move the heavy include to the corresponding .cpp (or wherever the type’s definition is actually required).
  src/paimon/core/mergetree/compact/compact_strategy.h:1
• Including bucketed_dv_maintainer.h in this header can significantly increase compile-time coupling. Since the API only needs std::shared_ptr<BucketedDvMaintainer>, you can forward-declare class BucketedDvMaintainer; in the paimon namespace and move the heavy include to the corresponding .cpp (or wherever the type’s definition is actually required).

---

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

[lucasfang]

+1