fix: include class UDF constructor state in cache identity - #1913
Open
atirna wants to merge 6 commits into
Open
Conversation
Differently configured Mapper/Aggregator instances were hashing as the same step, so the second call reused the first instance's cached rows.
atirna
force-pushed
the
fix/udf-hash-constructor-state
branch
from
August 17, 2026 05:58
5f86cbd to
4e736ec
Compare
Author
|
looks like there is competing work in progress, i am leaving it for them |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
I included class UDF constructor state in the UDF hash. Previously two Mapper/Aggregator instances that only differed in constructor args hashed the same, so the second call reused the first instance's cached rows. Now those instances get distinct cache keys. setup() runtime attrs (locks, loaded models) still dont move the hash until the UDF is rebound.
Fixes #1903
Why
Wrong answers that look plausible. The second configured UDF was served the first one's rows with no error or version bump.
Verification
main:CountAbove(0)thenCountAbove(3)both returned[(1, 5), (2, 5)]CountAbove(3)returns[(1, 2), (2, 2)]python -m pytest tests/unit/test_query_steps_hash.py tests/unit/test_datachain_hash.py tests/unit/lib/test_udf.py -q --benchmark-skip