[SPARK-52738][SQL] Support aggregating the TIME type with a UDAF when the underlying buffer is an UnsafeRow
#51430
+12
−8
Conversation
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
bersprockets
commented
Jul 9, 2025
| // Below we want to test with *both* UnsafeRow and safe row as the underlying | ||
| // buffer. | ||
| val mutableDataTypes = dataTypes.filter(UnsafeRow.isMutable) | ||
| Seq(dataTypes, mutableDataTypes).foreach { dataTypes => |
There was a problem hiding this comment.
Note that even without the change to BufferSetterGetterUtils, the unsafe buffer test will sometimes succeed. That's because the inappropriate setter function works fine for nulls and RandomDataGenerator will occasionally give us a null TIME value for the row where id == 50.
There was a problem hiding this comment.
Also, this test was succeeding even with TIME types because SPARK-41359 accidentally removed the unsafe buffer part of the test.
|
cc @MaxGekk |
MaxGekk
approved these changes
Jul 10, 2025
|
+1, LGTM. Merging to master. |
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Outdated suggestions cannot be applied.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.
What changes were proposed in this pull request?
BufferSetterGetterUtilsto useInternalRow.setLongfor setting TIME values rather thenInternalRow.update.BufferSetterGetterUtilsto useInternalRow.getLongfor getting TIME values.AggregationQuerySuiteso that it checks aggregation with both an unsafe and safe aggregation buffer. Since SPARK-41359, that test has been testing with only a safe aggregation buffer.Why are the changes needed?
When a query uses a UDAF to aggregate a TIME column , and all other columns are "mutable" (as determined by
UnsafeRow#isMutable), the aggregator creates anUnsafeRowfor the low-level aggregation buffer.However, the wrapper of that buffer (
MutableAggregationBufferImpl) fails to properly set up a field setter function for the TIME column, so it attempts to callUnsafeRow.updateon the underlying buffer. TheUnsafeRowinstance throwsorg.apache.spark.SparkUnsupportedOperationException:See SPARK-52738 for a reproduction example.
Does this PR introduce any user-facing change?
No. The TIME type is not released yet.
How was this patch tested?
Updated a unit test.
Was this patch authored or co-authored using generative AI tooling?
No.