Repository navigation
[Kernel connector] Implements SupportsRuntimeV2Filtering in SparkScan for DPP - #5269
Merged
Merged
Conversation
| } | ||
|
|
||
| assertEquals(new HashSet<>(expectedPartitionFilesAfterDpp), new HashSet<>(afterDppFiles)); | ||
| assertEquals(expectedTotalBytesAfterDpp, afterDppTotalBytes); |
Collaborator
There was a problem hiding this comment.
how does the check of total bytes work? Shall we check the number of filtered partitioned files as well?
Contributor
Author
There was a problem hiding this comment.
The expectedTotalBytesAfterDpp is calculated by traversing all partition files (i.e., beforeDppFiles in the code) and sum the ones containing the partition value segment in its path, which are also the ones satisfying the runtime filter.
For example, golden table has 5 partition files
- .../date=20180520/city=hz/part-00000-....snappy.parquet
- .../date=20180718/city=hz/part-00000-....snappy.parquet
- .../date=20180512/city=sh/part-00000-....snappy.parquet
- .../date=20180520/city=bj/part-00000-....snappy.parquet
- .../date=20181212/city=sz/part-00000-....snappy.parquet
To check the correctness for runtime filter city=hz, it traverses all partition files and finds the ones with city=hz in its path as the expected output.
Added size check.
…ession contains all predicates listed in org.apache.spark.sql.connector.expressions.filter.Predicate. (TODO: add white box tests for conversion)
| * Evaluate a predicate on partition values. | ||
| * Return true if the partition values satisfy the predicate, false otherwise. | ||
| */ | ||
| private boolean evaluatePartitionValueOnPredicate( |
Collaborator
There was a problem hiding this comment.
Let's have these code changes in Apache Spark
| * @return Catalyst Expression representing the resolved expression, or a literal true for | ||
| * unsupported expressions | ||
| */ | ||
| private static Expression resolveExpression( |
Collaborator
There was a problem hiding this comment.
how about dsv2ExpressionToCatalystExpression
| dsv2PredicateToCatalystExpression(predicate, partitionSchema); | ||
| if (catalystExpr.isPresent()) { | ||
| BasePredicate predicateEvaluator = | ||
| org.apache.spark.sql.catalyst.expressions.Predicate.create(catalystExpr.get()); |
Collaborator
There was a problem hiding this comment.
Since this always happens on driver, shall we call createInterpreted to avoid codegen?
gengliangwang
approved these changes
Oct 2, 2025
huashi-st
pushed a commit
to huashi-st/delta
that referenced
this pull request
Apr 24, 2026
… for DPP (delta-io#5269) <!-- Thanks for sending a pull request! Here are some tips for you: 1. If this is your first time, please read our contributor guidelines: https://github.com/delta-io/delta/blob/master/CONTRIBUTING.md 2. If the PR is unfinished, add '[WIP]' in your PR title, e.g., '[WIP] Your PR title ...'. 3. Be sure to keep the PR description updated to reflect all changes. 4. Please write your PR title to summarize what this PR proposes. 5. If possible, provide a concise example to reproduce the issue for a faster review. 6. If applicable, include the corresponding issue number in the PR title and link it in the body. --> #### Which Delta project/connector is this regarding? <!-- Please add the component selected below to the beginning of the pull request title For example: [Spark] Title of my pull request --> - [x] Spark - [ ] Standalone - [ ] Flink - [ ] Kernel - [ ] Other (fill in here) ## Description <!-- - Describe what this PR changes. - Describe why we need the change. If this PR resolves an issue be sure to include "Resolves #XXX" to correctly link and close the issue upon merge. --> This PR implements SupportsRuntimeV2Filtering for SparkScan of spark-kernel-dsv2, in order to enable dynamic partition pruning (i.e., DPP). ## How was this patch tested? <!-- If tests were added, say they were added here. Please make sure to test the changes thoroughly including negative and positive cases if possible. If the changes were tested in any way other than unit tests, please clarify how you tested step by step (ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future). If the changes were not tested, please explain why. --> Unit tests and golden table tests. ## Does this PR introduce _any_ user-facing changes? <!-- If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible. If possible, please also clarify if this is a user-facing change compared to the released Delta Lake versions or within the unreleased branches such as master. If no, write 'No'. --> No
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
2 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.
Which Delta project/connector is this regarding?
Description
This PR implements SupportsRuntimeV2Filtering for SparkScan of spark-kernel-dsv2, in order to enable dynamic partition pruning (i.e., DPP).
How was this patch tested?
Unit tests and golden table tests.
Does this PR introduce any user-facing changes?
No