[vortex] Preserve physical row order in scans - #9003
Open
huwh wants to merge 1 commit into
Open
Conversation
huwh
force-pushed
the
huwh/fix-vortex-scan-row-order
branch
2 times, most recently
from
August 3, 2026 12:24
fa65679 to
b6784b3
Compare
huwh
marked this pull request as ready for review
August 3, 2026 12:27
Author
|
@JingsongLi Hi, could you please take a look? Thanks! |
There was a problem hiding this comment.
Pull request overview
Note
Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.
This PR fixes a correctness issue where VortexRecordsReader relied on monotonically increasing returnedPosition() while the underlying Vortex scan could return partitions out of physical row order, potentially mismatching Vortex vector rows with main-file rows.
Changes:
- Force ordered Vortex scans by default in
VortexRecordsReader. - Add a regression test that validates physical row order is preserved across multiple scan partitions.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| paimon-vortex/paimon-vortex-format/src/main/java/org/apache/paimon/format/vortex/VortexRecordsReader.java | Requests ordered(true) on Vortex scans to preserve physical row order and align returnedPosition() with row IDs. |
| paimon-vortex/paimon-vortex-format/src/test/java/org/apache/paimon/format/vortex/VortexReaderWriterTest.java | Adds a regression test intended to demonstrate unordered scans can misorder results and ordered scans preserve physical row order. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+655
to
+659
| List<Integer> unorderedIds = readIds(testFile, selectedRows, false); | ||
| assertFalse( | ||
| expectedIds.equals(unorderedIds), | ||
| "The unordered scan unexpectedly preserved physical row order"); | ||
| assertEquals(expectedIds, readIds(testFile, selectedRows, true)); |
Comment on lines
+652
to
+682
| int previousWorkerCount = NativeRuntime.workerCount(); | ||
| NativeRuntime.setWorkerThreads(2); | ||
| try { | ||
| List<Integer> unorderedIds = readIds(testFile, selectedRows, false); | ||
| assertFalse( | ||
| expectedIds.equals(unorderedIds), | ||
| "The unordered scan unexpectedly preserved physical row order"); | ||
| assertEquals(expectedIds, readIds(testFile, selectedRows, true)); | ||
|
|
||
| try (VortexRecordsReader reader = | ||
| new VortexRecordsReader( | ||
| testFile, | ||
| rowType, | ||
| rowType, | ||
| selectedRows, | ||
| null, | ||
| Collections.emptyMap())) { | ||
| int readCount = 0; | ||
| FileRecordIterator<InternalRow> batch; | ||
| while ((batch = reader.readBatch()) != null) { | ||
| InternalRow row; | ||
| while ((row = batch.next()) != null) { | ||
| assertEquals(batch.returnedPosition(), row.getInt(0)); | ||
| readCount++; | ||
| } | ||
| } | ||
| assertEquals(selectedRows.length, readCount); | ||
| } | ||
| } finally { | ||
| NativeRuntime.setWorkerThreads(previousWorkerCount); | ||
| } |
Comment on lines
+723
to
+731
| private static String payload(int rowId) { | ||
| char[] chars = new char[4_096]; | ||
| int state = rowId + 1; | ||
| for (int i = 0; i < chars.length; i++) { | ||
| state = state * 1_103_515_245 + 12_345; | ||
| chars[i] = (char) ('a' + ((state >>> 16) & 15)); | ||
| } | ||
| return new String(chars); | ||
| } |
Vortex scans are unordered by default, while Paimon derives returned positions monotonically. Request ordered scans so dedicated Vortex files stay aligned with their corresponding main-file rows. Generated-by: OpenAI Codex
huwh
force-pushed
the
huwh/fix-vortex-scan-row-order
branch
from
August 3, 2026 13:08
b6784b3 to
fbebd5a
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Purpose
VortexRecordsReaderderivesreturnedPosition()from a monotonically increasing iterator, but the VortexDataSource/ScanAPI is unordered by default. If internal scan partitions complete out of order, dedicated Vortex vector rows can be paired with the wrong rows from the main data file.This change explicitly requests ordered Vortex scans and adds a deterministic test for the scan contract.
This PR is kept as a draft while waiting for issue assignment and maintainer feedback.
Fixes #9002.
Tests
mvn -pl paimon-vortex/paimon-vortex-format -am -Pfast-build -DfailIfNoTests=false -DwildcardSuites=none -Dtest=VortexRecordsReaderTest#testScanPreservesPhysicalRowOrder testexpected: <true> but was: <false>.mvn -pl paimon-vortex/paimon-vortex-format -am -Pfast-build -DfailIfNoTests=false -DwildcardSuites=none -Dtest=VortexRecordsReaderTest,VortexReaderWriterTest test(15 tests passed)mvn -pl paimon-vortex/paimon-vortex-format -am -DskipTests packagegit diff --check