Approximation framework to support numeric search_after queries
#18896
+414
−18
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
Signed-off-by: Prudhvi Godithi <[email protected]>
Signed-off-by: Prudhvi Godithi <[email protected]>
Signed-off-by: Prudhvi Godithi <[email protected]>
Signed-off-by: Prudhvi Godithi <[email protected]>
Signed-off-by: Prudhvi Godithi <[email protected]>
|
Adding @kkewwei @msfroh @harshavamsi to please take a look. |
prudhvigodithi
commented
Aug 2, 2025
| final byte[] effectiveLower = computeEffectiveLowerBound(); | ||
| final byte[] effectiveUpper = computeEffectiveUpperBound(); | ||
|
|
||
| this.pointRangeQuery = new PointRangeQuery( |
There was a problem hiding this comment.
I have to re-instantiate this PointRangeQuery as there is a logic in ApproximatePointRangeQuery.scorerSupplier to use pointRangeQueryWeight.scorerSupplier when the size requested is greater than the segment docs.
if (size > values.size()) {
return pointRangeQueryWeight.scorerSupplier(context);
}
Without the re-instantiate, it will continue to use the original range bounds and not the new computed search_after range bounds which leads to incorrect search results.
I'm open for thoughts here.
There was a problem hiding this comment.
The Lucene PointRangeQuery now uses ConstantScoreScorerSupplier which uses DenseConjunctionBulkScorer on a specific cost. Overall I see a code matchAll which can give good advantage. So looks like its good for us to have this if condition in ApproximatePointRangeQuery.
|
{"run-benchmark-test": "id_4"} |
|
{"run-benchmark-test": "id_13"} |
|
❌ Gradle check result for a68c544: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
|
The Jenkins job url is https://build.ci.opensearch.org/job/benchmark-pull-request/3963/ . Final results will be published once the job is completed. |
|
The Jenkins job url is https://build.ci.opensearch.org/job/benchmark-pull-request/3964/ . Final results will be published once the job is completed. |
Signed-off-by: Prudhvi Godithi <[email protected]>
|
❌ Gradle check result for f8e0670: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
Benchmark ResultsBenchmark Results for Job: https://build.ci.opensearch.org/job/benchmark-pull-request/3964/
|
Benchmark Baseline Comparison ResultsBenchmark Results for Job: https://build.ci.opensearch.org/job/benchmark-compare/139/
|
Signed-off-by: Prudhvi Godithi <[email protected]>
|
❌ Gradle check result for c5ddc30: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
Benchmark ResultsBenchmark Results for Job: https://build.ci.opensearch.org/job/benchmark-pull-request/3963/
|
Benchmark Baseline Comparison ResultsBenchmark Results for Job: https://build.ci.opensearch.org/job/benchmark-compare/140/
|
Signed-off-by: Prudhvi Godithi <[email protected]>
|
❌ Gradle check result for 2f95b6e: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
Signed-off-by: Prudhvi Godithi <[email protected]>
|
❌ Gradle check result for a7b5001: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
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.
This PR targets to support
search_afterqueries from the solution described in this issue #18546.ASCsort and upper bound forDESCsort. Example:Adds new
byte[] encodePoint(Object value, boolean roundUp);inNumericPointEncoderas todaysearch_aftervalue is returned as object and requires converting it into the appropriate field specific byte[].search_afterwithASCsort we want values greater than thesearch_aftervalue and forDESCsort we want values less than thesearch_after.Handles the search_after conversation for dates with date math rounding operations, here is the sample query. Note the existing
parseToLongis for date math rounding operations and not for creatingsearch_afterbounds.Coming from this PR https://github.com/opensearch-project/OpenSearch/pull/18763/files where approximation code path is disabled for multi sort, hence
search_aftershould work as expected when dealing with multiple tie breakersearch_aftervaluesRegarding benchmarks following are the queries we should see improvement from the existing workloads
desc_sort_with_after_timestamp,asc_sort_with_after_timestampdesc_sort_with_after_timestamp,asc_sort_with_after_timestampasc_sort_with_after_population,desc_sort_with_after_geonameid,asc_sort_with_after_geonameidDescription
[Describe what this change achieves]
Related Issues
Resolves #18546
Check List
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.