Skip to content

docs: MultiJoinStreamingDataset design plan (V1/V2) - #869

Merged
tchaton merged 10 commits into
mainfrom
docs/multi-join-streaming-dataset-plan
Aug 10, 2026
Merged

docs: MultiJoinStreamingDataset design plan (V1/V2)#869
tchaton merged 10 commits into
mainfrom
docs/multi-join-streaming-dataset-plan

Conversation

@tchaton

@tchaton tchaton commented Aug 10, 2026

Copy link
Copy Markdown
Collaborator

Summary

  • Adds a detailed design plan for MultiJoinStreamingDataset under .claude/plans/.
  • Describes Phase V1: strict aligned LitData chunks + validated wrapper over ParallelStreamingDataset.
  • Describes Phase V2: logical partitions with independently compacted column families (same public API).
  • Covers the polished writer/reader API, snapshot/manifest model, alignment invariants, migration workflow, tests, and risks.

This is a design document only — no runtime code changes.

Test plan

  • Review plan with customer / ML infra stakeholders
  • Confirm V1 vs V2 scope and delivery order
  • No code execution required for this PR

Made with Cursor

Customer-facing proposal for co-partitioned multi-table streaming with
an atomic version manifest, covering phase V1 aligned LitData chunks and
phase V2 independently compacted column families.

Co-authored-by: Cursor <cursoragent@cursor.com>
@tchaton
tchaton requested a review from justusschock as a code owner August 10, 2026 12:39
pre-commit-ci Bot and others added 4 commits August 10, 2026 12:39
Replace customer-specific entity names, paths, and examples with
generic multi-table terminology so the design can be shared publicly.

Co-authored-by: Cursor <cursoragent@cursor.com>
Snapshot IDs are human-readable publish names, not chunk content
hashes. Examples now use table_0 / table_1 / table_2 throughout.

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
@codecov-commenter

codecov-commenter commented Aug 10, 2026

Copy link
Copy Markdown

⚠️ Please install the 'codecov app svg image' to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81%. Comparing base (cf72e73) to head (78cbe82).
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files
@@         Coverage Diff         @@
##           main   #869   +/-   ##
===================================
  Coverage    81%    81%           
===================================
  Files        56     56           
  Lines      9303   9303           
===================================
  Hits       7563   7563           
  Misses     1740   1740           
🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

tchaton and others added 5 commits August 10, 2026 13:08
Drop MultiJoinWriter / key_name in favor of optimize-like
joint_optimize calls that can run as distributed jobs and
publish snapshots on success.

Co-authored-by: Cursor <cursoragent@cursor.com>
Lock join_optimize + join_update_optimize, input planning, sequential key validation, uniqueness, and snapshot concurrency rules into the design plan.

Co-authored-by: Cursor <cursoragent@cursor.com>
@tchaton
tchaton merged commit 8d786cc into main Aug 10, 2026
34 of 35 checks passed
@tchaton
tchaton deleted the docs/multi-join-streaming-dataset-plan branch August 10, 2026 15:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants