eejbyfeldt commented on code in PR #12095:
URL: https://github.com/apache/datafusion/pull/12095#discussion_r1730299634
##########
datafusion/functions-aggregate/src/variance.rs:
##########
@@ -344,3 +391,181 @@ impl Accumulator for VarianceAccumulator {
true
}
}
+
+#[derive(Debug)]
+pub struct VarianceGroupsAccumulator {
+ m2s: Vec<f64>,
+ means: Vec<f64>,
+ counts: Vec<u64>,
+ stats_type: StatsType,
+}
+
+impl VarianceGroupsAccumulator {
+ pub fn new(s_type: StatsType) -> Self {
+ Self {
+ m2s: Vec::new(),
+ means: Vec::new(),
+ counts: Vec::new(),
+ stats_type: s_type,
+ }
+ }
+
+ fn resize(&mut self, total_num_groups: usize) {
+ self.m2s.resize(total_num_groups, 0.0);
+ self.means.resize(total_num_groups, 0.0);
+ self.counts.resize(total_num_groups, 0);
+ }
+
+ fn merge<F>(
+ group_indices: &[usize],
+ counts: &UInt64Array,
+ means: &Float64Array,
+ m2s: &Float64Array,
+ opt_filter: Option<&BooleanArray>,
+ mut value_fn: F,
+ ) where
+ F: FnMut(usize, u64, f64, f64) + Send,
+ {
+ assert_eq!(counts.null_count(), 0);
+ assert_eq!(means.null_count(), 0);
+ assert_eq!(m2s.null_count(), 0);
+
+ match opt_filter {
+ None => {
+ group_indices
+ .iter()
+ .zip(counts.values().iter())
+ .zip(means.values().iter())
+ .zip(m2s.values().iter())
+ .for_each(|(((&group_index, &count), &mean), &m2)| {
+ value_fn(group_index, count, mean, m2);
+ });
+ }
+ Some(filter) => {
+ group_indices
+ .iter()
+ .zip(counts.values().iter())
+ .zip(means.values().iter())
+ .zip(m2s.values().iter())
+ .zip(filter.iter())
+ .for_each(
+ |((((&group_index, &count), &mean), &m2),
filter_value)| {
+ if let Some(true) = filter_value {
+ value_fn(group_index, count, mean, m2);
+ }
+ },
+ );
+ }
+ }
+ }
+
+ pub fn variance(
+ &mut self,
+ emit_to: datafusion_expr::EmitTo,
+ ) -> (Vec<f64>, NullBuffer) {
+ let mut counts = emit_to.take_needed(&mut self.counts);
+ let _ = emit_to.take_needed(&mut self.means);
Review Comment:
Yeah, there is probably no performance gains as it would need to do the same
work as take_needed.
I guess it might be worth adding if this becomes a common pattern in more
implementations.
--
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: github-unsubscr...@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org
