Fix strict metrics evaluator over-pruning files for NotEqualTo/NotIn with partial nulls#3521
Open
tanmayrauth wants to merge 1 commit into
Conversation
…with partial nulls
_StrictMetricsEvaluator.visit_not_equal and visit_not_in short-circuited on
_can_contain_nulls / _can_contain_nans (null/NaN count > 0) and returned
ROWS_MUST_MATCH without checking the value bounds. A file holding any null or
NaN was therefore reported as fully matching the predicate, even when a
non-null value inside the bounds did not match.
This drives _DeleteFiles (table/update/snapshot.py): ROWS_MUST_MATCH drops the
whole data file without rewriting it. So delete(NotEqualTo("x", 5)) against a
file with stats [null, 5] and bounds lower=upper=5 would delete the entire
file, silently losing the row with value 5 that should have survived.
Every other strict ROWS_MUST_MATCH path already guards on the "only" variants
(_contains_nulls_only / _contains_nans_only), matching the reference
StrictMetricsEvaluator. Switch both methods to the same guard so that an
all-null/all-NaN column still short-circuits to ROWS_MUST_MATCH (those rows
satisfy not-equal/not-in), while a partially-null column falls through to the
bounds check.
Update the existing NotIn-on-some-nulls test that encoded the buggy result and
add a regression test covering the [null, value] / bounds-include-literal case
for both NotEqualTo and NotIn.
Fixes apache#3498 (partially)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
_StrictMetricsEvaluator.visit_not_equal and visit_not_in short-circuited on _can_contain_nulls / _can_contain_nans (null/NaN count > 0) and returned ROWS_MUST_MATCH without checking the value bounds. A file holding any null or NaN was therefore reported as fully matching the predicate, even when a non-null value inside the bounds did not match.
This drives _DeleteFiles (table/update/snapshot.py): ROWS_MUST_MATCH drops the whole data file without rewriting it. So delete(NotEqualTo("x", 5)) against a file with stats [null, 5] and bounds lower=upper=5 would delete the entire file, silently losing the row with value 5 that should have survived.
Every other strict ROWS_MUST_MATCH path already guards on the "only" variants (_contains_nulls_only / _contains_nans_only), matching the reference StrictMetricsEvaluator. Switch both methods to the same guard so that an all-null/all-NaN column still short-circuits to ROWS_MUST_MATCH (those rows satisfy not-equal/not-in), while a partially-null column falls through to the bounds check.
Update the existing NotIn-on-some-nulls test that encoded the buggy result and add a regression test covering the [null, value] / bounds-include-literal case for both NotEqualTo and NotIn.
Fixes #3498 (partially)
Rationale for this change
Are these changes tested?
Are there any user-facing changes?