pepijnve opened a new pull request, #20105: URL: https://github.com/apache/datafusion/pull/20105
## Which issue does this PR close? - Related to #11570. ## Rationale for this change The `case_when` microbenchmark that covers the pattern `CASE WHEN d != 0 THEN n / d ELSE NULL END` pattern is parameterised over the percentage of zeroes in the `d` column. The benchmark uses the condition `d > 0` rather than `d != 0` though which is a bit misleading. In the '0% zeroes' run one would expect the else branch to never be taken, but because slightly less than 50% of the `d` values is negative, it's still taken 50% of the time. This PR adjust the benchmark to use `d != 0` instead. ## What changes are included in this PR? - Adjust the divide by zero benchmark to use `d != 0` as condition - Remove the duplicate benchmark, the div-by-zero variant is sufficient to compare changes across branches - Add a couple of SLTs to cover the `CASE` pattern ## Are these changes tested? Manual testing ## Are there any user-facing changes? No -- 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]
