having and count() estimation

englefly · Aug 12, 2024 · 6f649e7 · 6f649e7
1 parent 0520b2f
commit 6f649e7
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 10 deletions.
diff --git a/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/ExpressionEstimation.java b/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/ExpressionEstimation.java
@@ -373,7 +373,9 @@ public ColumnStatistic visitMax(Max max, Statistics context) {
     @Override
     public ColumnStatistic visitCount(Count count, Statistics context) {
         double width = count.getDataType().width();
-        return new ColumnStatisticBuilder().setCount(1D).setAvgSizeByte(width).setNumNulls(0)
+        // for scalar agg, ndv and row count will be normalized by 1 in StatsCalculator.computeAggregate()
+        return new ColumnStatisticBuilder().setNdv(context.getRowCount()).setCount(context.getRowCount())
+                .setAvgSizeByte(width).setNumNulls(0)
                 .setDataSize(width).setMinValue(0).setMaxValue(context.getRowCount())
                 .setMaxExpr(null).setMinExpr(null).build();
     }

diff --git a/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/FilterEstimation.java b/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/FilterEstimation.java
@@ -171,15 +171,15 @@ public Statistics visitComparisonPredicate(ComparisonPredicate cp, EstimationCon
                 }
                 return false;
             };
-            boolean leftAgg = left.anyMatch(containsAggSlot);
-            boolean rightAgg = right.anyMatch(containsAggSlot);
-            // It means this predicate appears in HAVING clause.
-            if (leftAgg || rightAgg) {
-                double rowCount = context.statistics.getRowCount();
-                double newRowCount = Math.max(rowCount * DEFAULT_HAVING_COEFFICIENT,
-                        Math.max(statsForLeft.ndv, statsForRight.ndv));
-                return context.statistics.withRowCount(newRowCount);
-            }
+            // boolean leftAgg = left.anyMatch(containsAggSlot);
+            // boolean rightAgg = right.anyMatch(containsAggSlot);
+            // // It means this predicate appears in HAVING clause.
+            // if (leftAgg || rightAgg) {
+            //     double rowCount = context.statistics.getRowCount();
+            //     double newRowCount = Math.max(rowCount * 0.5,
+            //             Math.max(statsForLeft.ndv, statsForRight.ndv));
+            //     return context.statistics.withRowCount(newRowCount);
+            // }
         }
         if (!left.isConstant() && !right.isConstant()) {
             return calculateWhenBothColumn(cp, context, statsForLeft, statsForRight);