fix: Handle empty FieldsData in reduce/rerank for requery scenario #44917

weiliu1031 · 2025-10-16T12:27:36Z

When requery optimization is enabled, search results contain IDs
but empty FieldsData. During reduce/rerank operations, if the
first shard has empty FieldsData while others have data,
PrepareResultFieldData initializes an empty array, causing
AppendFieldData to panic when accessing array indices.

Changes:

Find first non-empty FieldsData as template in 3 functions:
reduceAdvanceGroupBy, reduceSearchResultDataWithGroupBy,
reduceSearchResultDataNoGroupBy
Add length check before 2 AppendFieldData calls in reduce
functions to prevent panic
Improve newRerankOutputs to find first non-empty fieldData
using len(FieldsData) check instead of GetSizeOfIDs
Add length check in appendResult before AppendFieldData
Add comprehensive unit tests for empty and partial empty
FieldsData scenarios in both reduce and rerank functions

This fix handles both pure requery (all empty) and mixed
scenarios (some empty, some with data) without breaking normal
search flow. The key improvement is checking FieldsData length
directly rather than IDs, as requery may have IDs but empty
FieldsData.

sre-ci-robot · 2025-10-16T12:28:22Z

[ci-v2-notice]
Notice: We are gradually rolling out the new ci-v2 system.

Legacy CI jobs remain unaffected, you can just ignore ci-v2 if you don't want to run it.
Additional "ci-v2/*" checkers will run for this PR to ensure the new ci-v2 system is working as expected.
For tests that exist in both v1 and v2, passing in either system is considered PASS.

To rerun ci-v2 checks, comment with:

/ci-rerun-code-check // for ci-v2/code-check
/ci-rerun-build // for ci-v2/build
/ci-rerun-ut-integration // for ci-v2/ut-integration
/ci-rerun-ut-go // for ci-v2/ut-go
/ci-rerun-ut-cpp // for ci-v2/ut-cpp
/ci-rerun-ut // for all ci-v2/ut-integration, ci-v2/ut-go, ci-v2/ut-cpp

If you have any questions or requests, please contact @zhikunyao.

issue: milvus-io#44909 pr: milvus-io#44917 This change adds validation to filter out empty search results before reranking operations in both regular and advanced search scenarios. When QueryNode returns empty search results (with no IDs or fields), the rerank process would panic when trying to access result data. This happens in hybrid search scenarios where some sub-requests may return empty results. Changes include: - Add empty result filtering in reduceResults method - Add empty result filtering in PostExecute for advanced search - Add empty result filtering in all reduce utility functions (reduceAdvanceGroupBY, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy, rankSearchResultDataByGroup, rankSearchResultDataByPk) - Add unit tests to verify empty result handling This fix ensures empty results are filtered out before being passed to ranking functions, preventing the panic. Signed-off-by: Wei Liu <[email protected]>

mergify · 2025-10-16T13:04:52Z

@weiliu1031 cpu-e2e job failed, comment /run-cpu-e2e can trigger the job again.

codecov · 2025-10-16T13:12:31Z

Codecov Report

❌ Patch coverage is 66.66667% with 9 lines in your changes missing coverage. Please review.
✅ Project coverage is 76.94%. Comparing base (a3a28a4) to head (7bade09).
⚠️ Report is 15 commits behind head on master.

Files with missing lines	Patch %	Lines
internal/proxy/search_reduce_util.go	57.14%	6 Missing and 3 partials ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           master   #44917       +/-   ##
===========================================
- Coverage   83.90%   76.94%    -6.97%     
===========================================
  Files         507     1830     +1323     
  Lines       77920   285194   +207274     
===========================================
+ Hits        65378   219431   +154053     
- Misses      12542    58562    +46020     
- Partials        0     7201     +7201

Components	Coverage Δ
Client	`78.03% <ø> (∅)`
Core	`83.90% <ø> (ø)`
Go	`75.03% <75.00%> (∅)`

Files with missing lines	Coverage Δ
internal/util/function/rerank/util.go	`84.26% <100.00%> (ø)`
internal/proxy/search_reduce_util.go	`82.16% <57.14%> (ø)`

... and 1321 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

mergify · 2025-10-16T13:13:34Z

@weiliu1031 go-sdk check failed, comment rerun go-sdk can trigger the job again.

issue: milvus-io#44909 pr: milvus-io#44917 This change adds validation to filter out empty search results before reranking operations in both regular and advanced search scenarios. When QueryNode returns empty search results (with no IDs or fields), the rerank process would panic when trying to access result data. This happens in hybrid search scenarios where some sub-requests may return empty results. Changes include: - Add empty result filtering in reduceResults method - Add empty result filtering in PostExecute for advanced search - Add empty result filtering in all reduce utility functions (reduceAdvanceGroupBY, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy, rankSearchResultDataByGroup, rankSearchResultDataByPk) - Add unit tests to verify empty result handling This fix ensures empty results are filtered out before being passed to ranking functions, preventing the panic. Signed-off-by: Wei Liu <[email protected]>

mergify · 2025-10-20T08:07:53Z

@weiliu1031 cpu-e2e job failed, comment /run-cpu-e2e can trigger the job again.

mergify · 2025-10-20T08:20:08Z

@weiliu1031 go-sdk check failed, comment rerun go-sdk can trigger the job again.

issue: milvus-io#44909 When requery optimization is enabled, search results contain IDs but empty FieldsData. During reduce/rerank operations, if the first shard has empty FieldsData while others have data, PrepareResultFieldData initializes an empty array, causing AppendFieldData to panic when accessing array indices. Changes: - Find first non-empty FieldsData as template in 3 functions: reduceAdvanceGroupBy, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy - Add length check before 2 AppendFieldData calls in reduce functions to prevent panic - Improve newRerankOutputs to find first non-empty fieldData using len(FieldsData) check instead of GetSizeOfIDs - Add length check in appendResult before AppendFieldData - Add comprehensive unit tests for empty and partial empty FieldsData scenarios in both reduce and rerank functions This fix handles both pure requery (all empty) and mixed scenarios (some empty, some with data) without breaking normal search flow. The key improvement is checking FieldsData length directly rather than IDs, as requery may have IDs but empty FieldsData. Signed-off-by: Wei Liu <[email protected]>

congqixia

/lgtm

sre-ci-robot · 2025-10-21T11:41:23Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: congqixia, weiliu1031

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [congqixia]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

…rio (#44919) issue: #44909 pr: #44917 When requery optimization is enabled, search results contain IDs but empty FieldsData. During reduce/rerank operations, if the first shard has empty FieldsData while others have data, PrepareResultFieldData initializes an empty array, causing AppendFieldData to panic when accessing array indices. Changes: - Find first non-empty FieldsData as template in 5 functions: reduceAdvanceGroupBY, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy, rankSearchResultDataByGroup, rankSearchResultDataByPk - Add length check before 4 AppendFieldData calls to prevent panic - Add unit tests for empty and partial empty FieldsData scenarios This fix handles both pure requery (all empty) and mixed scenarios (some empty, some with data) without breaking normal search flow. Signed-off-by: Wei Liu <[email protected]>

…uery scenario (milvus-io#44919) issue: milvus-io#44909 pr: milvus-io#44917 When requery optimization is enabled, search results contain IDs but empty FieldsData. During reduce/rerank operations, if the first shard has empty FieldsData while others have data, PrepareResultFieldData initializes an empty array, causing AppendFieldData to panic when accessing array indices. Changes: - Find first non-empty FieldsData as template in 5 functions: reduceAdvanceGroupBY, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy, rankSearchResultDataByGroup, rankSearchResultDataByPk - Add length check before 4 AppendFieldData calls to prevent panic - Add unit tests for empty and partial empty FieldsData scenarios This fix handles both pure requery (all empty) and mixed scenarios (some empty, some with data) without breaking normal search flow. Signed-off-by: Wei Liu <[email protected]>

…uery scenario (#44919) (#45016) issue: #44909 pr: #44917 When requery optimization is enabled, search results contain IDs but empty FieldsData. During reduce/rerank operations, if the first shard has empty FieldsData while others have data, PrepareResultFieldData initializes an empty array, causing AppendFieldData to panic when accessing array indices. Changes: - Find first non-empty FieldsData as template in 5 functions: reduceAdvanceGroupBY, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy, rankSearchResultDataByGroup, rankSearchResultDataByPk - Add length check before 4 AppendFieldData calls to prevent panic - Add unit tests for empty and partial empty FieldsData scenarios This fix handles both pure requery (all empty) and mixed scenarios (some empty, some with data) without breaking normal search flow. Signed-off-by: Wei Liu <[email protected]>

…ilvus-io#44917) issue: milvus-io#44909 When requery optimization is enabled, search results contain IDs but empty FieldsData. During reduce/rerank operations, if the first shard has empty FieldsData while others have data, PrepareResultFieldData initializes an empty array, causing AppendFieldData to panic when accessing array indices. Changes: - Find first non-empty FieldsData as template in 3 functions: reduceAdvanceGroupBy, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy - Add length check before 2 AppendFieldData calls in reduce functions to prevent panic - Improve newRerankOutputs to find first non-empty fieldData using len(FieldsData) check instead of GetSizeOfIDs - Add length check in appendResult before AppendFieldData - Add comprehensive unit tests for empty and partial empty FieldsData scenarios in both reduce and rerank functions This fix handles both pure requery (all empty) and mixed scenarios (some empty, some with data) without breaking normal search flow. The key improvement is checking FieldsData length directly rather than IDs, as requery may have IDs but empty FieldsData. Signed-off-by: Wei Liu <[email protected]>

…44917) (#45137) issue: #44909 pr: #44917 When requery optimization is enabled, search results contain IDs but empty FieldsData. During reduce/rerank operations, if the first shard has empty FieldsData while others have data, PrepareResultFieldData initializes an empty array, causing AppendFieldData to panic when accessing array indices. Changes: - Find first non-empty FieldsData as template in 3 functions: reduceAdvanceGroupBy, reduceSearchResultDataWithGroupBy, reduceSearchResultDataNoGroupBy - Add length check before 2 AppendFieldData calls in reduce functions to prevent panic - Improve newRerankOutputs to find first non-empty fieldData using len(FieldsData) check instead of GetSizeOfIDs - Add length check in appendResult before AppendFieldData - Add comprehensive unit tests for empty and partial empty FieldsData scenarios in both reduce and rerank functions This fix handles both pure requery (all empty) and mixed scenarios (some empty, some with data) without breaking normal search flow. The key improvement is checking FieldsData length directly rather than IDs, as requery may have IDs but empty FieldsData. Signed-off-by: Wei Liu <[email protected]>

sre-ci-robot added the size/L Denotes a PR that changes 100-499 lines. label Oct 16, 2025

sre-ci-robot requested review from czs007 and yanliang567 October 16, 2025 12:27

mergify bot added dco-passed DCO check passed. kind/bug Issues or changes related a bug labels Oct 16, 2025

weiliu1031 mentioned this pull request Oct 16, 2025

fix: [2.5] Handle empty FieldsData in reduce/rerank for requery scenario #44919

Merged

weiliu1031 force-pushed the fix_rerank_panic branch 3 times, most recently from 8dae031 to 4e8fb8c Compare October 17, 2025 06:38

weiliu1031 changed the title ~~fix: Filter empty search results to prevent rerank panic~~ fix: fix: Handle empty FieldsData in reduce/rerank for requery scenario Oct 20, 2025

weiliu1031 changed the title ~~fix: fix: Handle empty FieldsData in reduce/rerank for requery scenario~~ fix: Handle empty FieldsData in reduce/rerank for requery scenario Oct 20, 2025

weiliu1031 force-pushed the fix_rerank_panic branch from 4e8fb8c to a14f83e Compare October 20, 2025 08:01

weiliu1031 force-pushed the fix_rerank_panic branch from a14f83e to 7bade09 Compare October 20, 2025 11:44

mergify bot added the ci-passed label Oct 20, 2025

congqixia approved these changes Oct 21, 2025

View reviewed changes

sre-ci-robot assigned congqixia Oct 21, 2025

sre-ci-robot added the lgtm label Oct 21, 2025

sre-ci-robot added the approved label Oct 21, 2025

sre-ci-robot merged commit 9dddc38 into milvus-io:master Oct 21, 2025
15 of 18 checks passed

weiliu1031 mentioned this pull request Oct 21, 2025

fix: [hotfix-2.5.19] Handle empty FieldsData in reduce/rerank for requery scenario (#44919) #45016

Merged

congqixia mentioned this pull request Oct 29, 2025

[Bug]: run HybridSearch but the service panic #45135

Open

1 task

weiliu1031 mentioned this pull request Oct 29, 2025

fix: Handle empty FieldsData in reduce/rerank for requery scenario (#44917) #45137

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix: Handle empty FieldsData in reduce/rerank for requery scenario #44917

fix: Handle empty FieldsData in reduce/rerank for requery scenario #44917

Uh oh!

weiliu1031 commented Oct 16, 2025 •

edited

Loading

Uh oh!

sre-ci-robot commented Oct 16, 2025

Uh oh!

mergify bot commented Oct 16, 2025

Uh oh!

codecov bot commented Oct 16, 2025 •

edited

Loading

Uh oh!

mergify bot commented Oct 16, 2025

Uh oh!

mergify bot commented Oct 20, 2025

Uh oh!

mergify bot commented Oct 20, 2025

Uh oh!

congqixia left a comment

Uh oh!

sre-ci-robot commented Oct 21, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

fix: Handle empty FieldsData in reduce/rerank for requery scenario #44917

fix: Handle empty FieldsData in reduce/rerank for requery scenario #44917

Uh oh!

Conversation

weiliu1031 commented Oct 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

sre-ci-robot commented Oct 16, 2025

Uh oh!

mergify bot commented Oct 16, 2025

Uh oh!

codecov bot commented Oct 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

mergify bot commented Oct 16, 2025

Uh oh!

mergify bot commented Oct 20, 2025

Uh oh!

mergify bot commented Oct 20, 2025

Uh oh!

congqixia left a comment

Choose a reason for hiding this comment

Uh oh!

sre-ci-robot commented Oct 21, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

weiliu1031 commented Oct 16, 2025 •

edited

Loading

codecov bot commented Oct 16, 2025 •

edited

Loading