Do not use Min or Max as Top's surrogate when there is an outputField #138380
+98
−4
Conversation
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
ncordon
added
>bug
Team:Analytics
Collaborator
|
Hi @ncordon, I've created a changelog YAML for you. |
ncordon
force-pushed
the
fix-edge-case-in-top
branch
from
cd95fe8 to
954c27f
Compare
Collaborator
|
Pinging @elastic/es-analytical-engine (Team:Analytics) |
Collaborator
|
Hi @ncordon, I've created a changelog YAML for you. |
ivancea
reviewed
Nov 20, 2025
| } | ||
| if (orderField() instanceof Literal && limitField() instanceof Literal && limitValue() == 1) { | ||
| // To replace Top by Min or Max, we cannot have an `outputField` | ||
| if (orderField() instanceof Literal && limitField() instanceof Literal && limitValue() == 1 && outputField() == null) { |
Contributor
There was a problem hiding this comment.
Can we add some CSV test reproducing this case? To ensure that this indeed failed before the fix.
A surrogate changing the result type of the expression is a quite hardcore error at planning time.
Our unit tests happen to execute the surrogate logic, but it's really executed as part of the planning, so having an integration test checking it looks safer (Apart of the unit tests, which look fine)
Contributor
Author
There was a problem hiding this comment.
Done. If I backtrack the fix, the following error is thrown in those tests (which I find not very helpful), so you are right the planner would not let this through:
Output has changed from [[shortest_employee{r}#577, tallest_employee{r}#581]] to [[shortest_employee{r}#577, tallest_employee{r}#581]].
szybia
added a commit
to szybia/elasticsearch
that referenced
this pull request
Nov 24, 2025
…-json * upstream/main: (247 commits) Mute org.elasticsearch.xpack.inference.integration.SemanticTextIndexOptionsIT testValidateIndexOptionsWithBasicLicense elastic#138513 Mute org.elasticsearch.xpack.esql.heap_attack.HeapAttackLookupJoinIT testLookupExplosionBigString elastic#138510 This shouldn't be zero (elastic#138501) sum of empty histogram is now null (elastic#138378) Test ES|QL bfloat16 support (elastic#138499) Fix exception handling in S3 `compareAndExchangeRegister` (elastic#138488) Mute org.elasticsearch.xpack.exponentialhistogram.ExponentialHistogramFieldMapperTests testFormattedDocValues elastic#138504 Mute org.elasticsearch.ingest.geoip.IngestGeoIpClientYamlTestSuiteIT test {yaml=ingest_geoip/60_ip_location_databases/Test adding, getting, and removing ip location databases} elastic#138502 ESQL: Refactor HeapAttackIT (elastic#138432) [Inference API] Add ElasticInferenceServiceDenseTextEmbeddingsServiceSettings to InferenceNamedWriteablesProvider (elastic#138484) Store split indices (elastic#138396) ES|QL Update CHUNK to support chunking_settings as optional argument (elastic#138123) Extract common blob-update logic in `S3HttpHandler` (elastic#138490) Cleanup esql request building api (elastic#138398) Round sum and avg in exponential_histogram CSV tests (elastic#138472) ESQL: load exponential_histogram total count as double instead of long (elastic#138417) [SIMD] Use fixed width native types for better Java interoperability (elastic#138429) Do not use Min or Max as Top's surrogate when there is an outputField (elastic#138380) ES|QL: Fix generative tests (elastic#138478) Mute org.elasticsearch.xpack.inference.integration.AuthorizationTaskExecutorIT testCreatesEisChatCompletion_DoesNotRemoveEndpointWhenNoLongerAuthorized elastic#138480 ...
afoucret
pushed a commit
to afoucret/elasticsearch
that referenced
this pull request
Nov 26, 2025
ncordon
added a commit
to ncordon/elasticsearch
that referenced
this pull request
Nov 26, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Why
Now that we have added
outputFieldtoTop, we should not be usingMaxorMinwhen there is anoutputFieldbecause those expressions do not take anoutputField.This was causing some test flakiness. The tests were not failing consistently because this only happens for a limit of 1 (and limit is also randomized in the tests).
Closes #134083