cloud-fan commented on code in PR #56382:
URL: https://github.com/apache/spark/pull/56382#discussion_r3432195056
##########
sql/core/src/test/scala/org/apache/spark/sql/DataFrameAggregateSuite.scala:
##########
@@ -1091,6 +1091,35 @@ class DataFrameAggregateSuite extends SharedSparkSession
)
}
+ test("SPARK-57329: mode normalizes -0.0/0.0 in the frequency buffer") {
Review Comment:
Coverage here is solid — top-level Double, top-level Float, and
`Array[Double]` for both `mode` and `pandas_mode`, and each case genuinely
distinguishes fixed from buggy behavior.
Two paths are exercised only indirectly and could be worth pinning down
(non-blocking):
- A struct or map complex type — same recursive
`NormalizeFloatingNumbers.normalize` path as the array case, but a different
shape.
- A NaN case — the complex-type lambda canonicalizes NaN bits, so this PR
also fixes NaN-in-complex-type collapse, but no test currently locks that in.
Low value (shared code path; differing NaN bit-patterns are a rare edge), so
purely your call.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
