GH-51684: [C++][Parquet] Keep the CDC chunking state across row groups

[kszucs]

Rationale for this change

The content-defined chunker was owned by the column writer, which is recreated for every row group, so the chunking state was reset at each row group boundary and page boundaries depended on where a row group starts. See #51684.

What changes are included in this PR?

• The file writer creates one chunker per leaf column when it opens and passes it to the column writers of every row group through a private ColumnWriter::Make() overload.
• The file writer computes each column's level information once and shares it with the chunker and the column writers.
• A column writer requires a chunker when content-defined chunking is enabled, so the public ColumnWriter::Make() now rejects such properties.
• NextColumn() advances the column ordinal together with the column metadata, before closing the previous column writer. Previously, if that close threw and the caller called NextColumn() again, the ordinal fell one column behind the metadata, and the new column writer would have used the previous column's chunker.

Are these changes tested?

Yes. IndependentOfRowGroupBoundaries writes the same table into one and into multiple row groups and expects the same page boundaries apart from the row group ends; the first commit adds it alone, failing without the fix.

Are there any user-facing changes?

Files with multiple row groups get different page boundaries than before, since the chunking now continues across row groups. ColumnWriter::Make() throws if the writer properties enable content-defined chunking; such column writers have to come from the file writer. Multi-threaded writes with content-defined chunking are about 1.7× slower until the follow-up #51692 keeps the chunker's state in registers (9.9 ms vs 5.8 ms writing 12 int32 columns of 1M rows on an Apple M4 Max, 4.4 ms with the follow-up): the chunkers are now created together, so the threads writing neighbouring columns contend for their cache lines.

Was AI used for this PR?

In accordance to the AI generation guidelines, please disclose below whether and how AI was used in this PR.

Claude Code wrote the code, tests and description under human direction.

PR code and description written by:

• Human
• AI

Reviewed before submission by:

• Human
• AI
• Not reviewed

🤖 Generated with Claude Code

• GitHub Issue: [C++][Parquet] Content-defined chunking resets its state at every row group #51684

[Copilot]

Copilot review overview

🟡 Changes recommended

The changed public factory breaks ABI and disables CDC for existing direct callers.

Review effort: Balanced
Findings: 1

Open (1)

• Preserve Make ABI and content-defined chunker creation · New

What changed in this PR

Persists Parquet content-defined chunking state across row groups.

Changes:

• Moves chunker ownership to the file writer.
• Passes chunkers into successive column writers.
• Adds row-group boundary independence coverage.

File	Description
cpp/​src/​parquet/​file_writer.cc	Manages per-column chunkers.
cpp/​src/​parquet/​column_writer.h	Extends the column-writer factory.
cpp/​src/​parquet/​column_writer.cc	Uses externally managed chunkers.
cpp/​src/​parquet/​chunker_internal.h	Updates ownership documentation.
cpp/​src/​parquet/​chunker_internal.cc	Stores level metadata by value.
cpp/​src/​parquet/​chunker_internal_test.cc	Tests stable page boundaries.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.