dwsmith1983 commented on PR #6712: URL: https://github.com/apache/datafusion-comet/pull/6712#issuecomment-6044706159
> Could we add a SQL file test for it, something like `expressions/math/round_nondeterministic_child.sql` with `-- Config: spark.sql.ansi.enabled=false`, an 8-row table and that query? Added in 1bfca75f3 as `expressions/math/round_nondeterministic_child.sql`, with ANSI off, an 8-row single-file table and your query. It uses `expect_dispatch(round)`, so it also fails if `round` over a double stops going through the dispatcher. On `main` it fails the way you saw: `a` matches Spark and `b` continues `a`'s sequence (0.64 then 0.39 in the first row, where Spark has 0.64 twice). It passes on this branch on Spark 4.1 and 3.5. -- 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]
