viirya opened a new issue, #23955: URL: https://github.com/apache/datafusion/issues/23955
### Is your feature request related to a problem or challenge? `PrimitiveDistinctCountAccumulator::update_batch` inserts values with a per-element `Option` check (`if let Some(value) = value`) on every row, even when the input array has no nulls. This is the hot path for `COUNT(DISTINCT <int/date/time/decimal>)`. The same accumulator's `merge_batch` already inserts wholesale from the values buffer (`self.values.extend(list.values())`), so the fast path is available — it's just not applied on the update path. ### Describe the solution you'd like When `array.null_count() == 0`, insert directly from the values buffer, skipping the per-element validity branch — mirroring `merge_batch` and the null-free fast paths recently added to `percentile_cont`/`median` in #23954. Measured ~66% faster on the null-free `count_distinct` benchmark. -- 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]
