Short-circuit BoolArray null count (#1509)

I believe this is the regression from 10a1f21 It checks both true_count and null_count, but the null count was triggering a full stats compute.
spiraldb · Nov 28, 2024 · 61e838e · 61e838e
1 parent 787106c
commit 61e838e
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 4 deletions.
diff --git a/vortex-array/src/array/bool/stats.rs b/vortex-array/src/array/bool/stats.rs
@@ -43,11 +43,12 @@ impl StatisticsVTable<NullableBools<'_>> for BoolEncoding {
         // Fast-path if we just want the true-count
         if matches!(
             stat,
-            Stat::TrueCount | Stat::Min | Stat::Max | Stat::IsConstant
+            Stat::TrueCount | Stat::Min | Stat::Max | Stat::IsConstant | Stat::NullCount
         ) {
+            let _null_count = array.1.count_set_bits();
             return Ok(StatsSet::bools_with_true_and_null_count(
                 array.0.bitand(array.1).count_set_bits(),
-                array.1.count_set_bits(),
+                array.1.len() - array.1.count_set_bits(),
                 array.0.len(),
             ));
         }
@@ -85,7 +86,7 @@ impl StatisticsVTable<BooleanBuffer> for BoolEncoding {
         // Fast-path if we just want the true-count
         if matches!(
             stat,
-            Stat::TrueCount | Stat::Min | Stat::Max | Stat::IsConstant
+            Stat::TrueCount | Stat::Min | Stat::Max | Stat::IsConstant | Stat::NullCount
         ) {
             return Ok(StatsSet::bools_with_true_and_null_count(
                 buffer.count_set_bits(),

diff --git a/vortex-array/src/stats/statsset.rs b/vortex-array/src/stats/statsset.rs
@@ -90,9 +90,10 @@ impl StatsSet {
         true_count: usize,
         null_count: usize,
         len: usize,
-    ) -> StatsSet {
+    ) -> Self {
         StatsSet::from_iter([
             (Stat::TrueCount, true_count.into()),
+            (Stat::NullCount, null_count.into()),
             (Stat::Min, (true_count == len).into()),
             (Stat::Max, (true_count > 0).into()),
             (