From 397992124a02da635fd42ec213bbb55e005eede0 Mon Sep 17 00:00:00 2001
From: xunxunmimi5577 <52647492+xunxunmimi5577@users.noreply.github.com>
Date: Thu, 25 Dec 2025 14:43:53 +0800
Subject: [PATCH] Make filer estimation more accurate with histogram

Make filer estimation more accurate with histogram
---
 .../plans/logical/statsEstimation/EstimationUtils.scala     | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/EstimationUtils.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/EstimationUtils.scala
index 7083014f1f38..8254827884f3 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/EstimationUtils.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/EstimationUtils.scala
@@ -243,10 +243,12 @@ object EstimationUtils {
     } else {
       // Computes the occupied portion of bins of the upperBound and lowerBound.
       val lowerBin = bins(lowerBinIndex)
-      val lowerPart = binHoldingRangePossibility(lowerBin.hi, lowerBound, lowerBin)
+      val lowerPart = if (lowerBin.hi == lowerBound && !lowerBoundInclusive) 0.0
+        else binHoldingRangePossibility(lowerBin.hi, lowerBound, lowerBin)
 
       val higherBin = bins(upperBinIndex)
-      val higherPart = binHoldingRangePossibility(upperBound, higherBin.lo, higherBin)
+      val higherPart = if (higherBin.lo == upperBound && !upperBoundInclusive) 0.0
+        else binHoldingRangePossibility(upperBound, higherBin.lo, higherBin)
 
       // The total number of bins is lowerPart + higherPart + bins between them
       lowerPart + higherPart + upperBinIndex - lowerBinIndex - 1