ANSI check for aggregates #3597
Merged
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: Alessandro Bellina <abellina@nvidia.com>
|
build |
|
Note I didn't include window agg tests, but I can certainly add some. I did verify that a simple window op using a sum, would fallback given the ANSI flag. |
revans2
reviewed
Sep 22, 2021
| @pytest.mark.parametrize('ansi_enabled', ['true', 'false']) | ||
| def test_hash_grpby_avg_nulls_ansi(data_gen, conf, ansi_enabled): | ||
| local_conf = copy_and_update(conf, {'spark.sql.ansi.enabled': ansi_enabled}) | ||
| assert_gpu_and_cpu_are_equal_collect( |
There was a problem hiding this comment.
I m confused why want to combine this into a single test. Doing it this way makes it so we cannot verify that we did fallback for Average and it also means that we explicitly allow things to not be on the GPU when we know that they should be in non-ansi mode.
There was a problem hiding this comment.
Sorry, I moved that test down together with the other ANSI stuff, I didn't change it. Will split it, totally agree.
There was a problem hiding this comment.
Done here: 4ddbb6b
revans2
approved these changes
Sep 22, 2021
|
build |
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Outdated suggestions cannot be applied.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.
Adds fallback for aggregates if ANSI mode is enabled. Specifically:
Sum,Count, andAverageall fallback in various ways.SumandCountcheck the types that they are producing (resultType), whereAverageunconditionally falls back, because it usesLonginternally. The other aggregates do not fallback.I am finding that adding
DecimalGento the mix for the testing is probably not possible. Spark is either optimizing it out (UnscaledValue), or we don't support the precision. But I could have missed something.This does not add the
CentralMomentAggvariety to the checks, since they are not merged yet, but those definitely need to fallback.