Fix NaN handling in GpuLessThanOrEqual and GpuGreaterThanOrEqual #9780
Conversation
This file contains 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: Jason Lowe <[email protected]>
|
build |
|
build |
1 similar comment
|
build |
|
build |
revans2
approved these changes
Nov 20, 2023
| @@ -156,6 +156,7 @@ def test_lte(data_gen): | |||
| lambda spark : binary_op_df(spark, data_gen).select( | |||
There was a problem hiding this comment.
Do we want some tests with hard coded NaN in them?
f.lit("NaN").cast(data_type)
Similar to what we do for None? Just so we know the error case is covered instead of waiting for it to possibly happen.
| @@ -258,7 +258,6 @@ def test_lt(data_descr): | |||
| s2 < f.col('b'), | |||
| f.col('a') < f.col('b'))) | |||
|
|
|||
| @datagen_overrides(seed=0, reason='https://github.com/NVIDIA/spark-rapids/issues/9711') | |||
| @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn) | |||
| def test_lte(data_descr): | |||
There was a problem hiding this comment.
Like in the other tests. Do we want to explicitly have a NaN Scalar in the tests to verify? Or a dedicated test for it?
There was a problem hiding this comment.
We're currently generating NaN and null often enough as special cases that this should be covered pretty well across multiple runs. Want to get these tests fixed for CI, can file followup if desired.
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Outdated suggestions cannot be applied.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.
Fixes #9751. Fixes #9711.
ColumnView.isNandoes not return any nulls (null input rows return false output rows), so the logic for handling a scalar NaN in GpuLessThanOrEqual and GpuGreaterThanOrEqual was accidentally clobbering nulls in the output. Updated to simply apply the validity vector of the input to the column vector full of true rows to produce the proper result for this corner case.