[refactor](storage) introduce the block transform chain by csun5285 · Pull Request #66528 · apache/doris

csun5285 · 2026-08-06T04:32:35Z

Third PR of the #64674 split stack (#65492, #66151). Adds storage/transform/block_transform.{h,cpp}: an immutable BlockTransformChain applied to every flushed block at the two segment_creator seams before the block reaches a segment writer.

What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary:

Release note

None

Check List (For Author)

Test
- Regression test
- Unit Test
- Manual test (add detailed scripts or steps below)
- No need to test or manual test. Explain why:
  - This is a refactor/code format and no logic has been changed.
  - Previous test can cover this change.
  - No code files have been changed.
  - Other reason
Behavior changed:
- No.
- Yes.
Does this need documentation?
- No.
- Yes.

Check List (For Reviewer who merge this PR)

Confirm the release note
Confirm test cases
Confirm document
Add branch pick label

…te/VariantParse/RowStoreFill at the segment_creator seams Third PR of the apache#64674 split stack (apache#65492, apache#66151). Adds storage/transform/block_transform.{h,cpp}: an immutable BlockTransformChain applied to every flushed block at the two segment_creator seams before the block reaches a segment writer. - ValidateStage: the schema/width checks both writers duplicated. - VariantParseStage: the non-partial-update variant parse both writers called. - RowStoreFillStage: registers a RowStoreColumnGenerator over a COW snapshot taken before variant parse, so the row-store column is still built from the raw variant representation. The vertical writer pumps it in bounded batches (_append_generated_column, 4MB / num_rows_per_block caps); the horizontal writer materializes it up front. Bridges removed by later PRs of the stack: binlog sub-writers get an empty chain (RowBinlogSegmentWriter still derives rows itself); partial update loads get [Validate] only (the writers still own the PU fill/parse/row-store work). Every block goes through exactly one of chain-or-writer per transform. New guards on previously crashing paths: PU without tablet context -> NotSupported; PU flushed through add_block (no segment id) -> InternalError. add_block never carries PU loads (callers: push handler, merger, schema change). Equivalence pinned by the golden segment-format tests (apache#65977): RowStoreAndSegmentCreatorPathsKeepTheirSegmentBytes and PartialUpdateAndRowBinlogPathsKeepTheirSegmentBytes cover the touched paths. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

hello-stephen · 2026-08-06T04:32:41Z

Thank you for your contribution to Apache Doris.
Don't know what should be done next? See How to process your PR.

Please clearly describe your PR:

What problem was fixed (it's best to include specific error reporting information). How it was fixed.
Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
What features were added. Why was this function added?
Which code was refactored and why was this part of the code refactored?
Which functions were optimized and what is the difference before and after the optimization?

csun5285 · 2026-08-06T04:33:48Z

run buildall

csun5285 · 2026-08-06T09:22:32Z

run buildall

hello-stephen · 2026-08-06T10:32:51Z

BE UT Coverage Report

Increment line coverage 94.32% (216/229) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	59.79% (26096/43643)
Line Coverage	44.11% (264593/599824)
Region Coverage	39.86% (210742/528735)
Branch Coverage	41.28% (96494/233768)

github-actions · 2026-08-06T10:35:19Z

PR approved by at least one committer and no changes requested.

github-actions · 2026-08-06T10:35:22Z

PR approved by anyone and no changes requested.

hello-stephen · 2026-08-06T10:36:26Z

BE Regression && UT Coverage Report