From fdc4602992328e75aefc1f5654f1ea2b998bd1ce Mon Sep 17 00:00:00 2001
From: "ZizhuanLiu(X-MAN)" <44973863@qq.com>
Date: Thu, 10 Sep 2026 20:45:45 +0800
Subject: [PATCH v5] Fix var_eq_const: sum selectivity of all matching MCV
 entries for nondeterministic collation

When evaluating selectivity for var_eq_const(), if the query uses a nondeterministic
collation which differs from the statistics collation, a single constant may match
multiple entries in the MCV list. The original code would stop at the first matched
MCV entry and return its frequency, leading to underestimated selectivity and poor-quality
query plans.

Enable full-MCV-array scan under the correct condition:
  the collation used for the comparison is nondeterministic and differs from the
	statistics collation.

When full scan is enabled, iterate over all MCV entries and accumulate the total
selectivity of every matching entry. Keep the fast-path early-break for all other
cases to avoid performance regression.

Discussion: https://www.postgresql.org/message-id/flat/tencent_A15A9D89A86F2E4B086ABA462578B9B64307@qq.com
Commitfest: https://commitfest.postgresql.org/patch/7075/
---
 src/backend/utils/adt/selfuncs.c | 56 ++++++++++++++++++++++++--------
 1 file changed, 43 insertions(+), 13 deletions(-)

diff --git a/src/backend/utils/adt/selfuncs.c b/src/backend/utils/adt/selfuncs.c
index e27ec9e5..0bd313c8 100644
--- a/src/backend/utils/adt/selfuncs.c
+++ b/src/backend/utils/adt/selfuncs.c
@@ -413,6 +413,7 @@ var_eq_const(VariableStatData *vardata, Oid oproid, Oid collation,
 		AttStatsSlot sslot;
 		bool		match = false;
 		int			i;
+		double		sumcommon = 0.0;
 
 		/*
 		 * Is the constant "=" to any of the column's most common values?
@@ -427,6 +428,7 @@ var_eq_const(VariableStatData *vardata, Oid oproid, Oid collation,
 		{
 			LOCAL_FCINFO(fcinfo, 2);
 			FmgrInfo	eqproc;
+			bool		scan_entire_mcv = false;
 
 			fmgr_info(opfuncoid, &eqproc);
 
@@ -446,6 +448,33 @@ var_eq_const(VariableStatData *vardata, Oid oproid, Oid collation,
 			else
 				fcinfo->args[0].value = constval;
 
+			/*
+			 * Scanning the entire MCV array is needed when the collation used
+			 * for the comparison is nondeterministic and differs from the
+			 * statistics collation. In this case, the comparison may match
+			 * multiple MCV values, so we must continue scanning after finding
+			 * a match.
+			 */
+			if (sslot.stacoll != collation && OidIsValid(collation))
+			{
+				pg_locale_t mylocale = pg_newlocale_from_collation(collation);
+				scan_entire_mcv = !mylocale->deterministic;
+			}
+
+			/*
+			 * Compare the constant expression with the MCVs.
+			 *
+			 * If the constant matches the current MCV, stop here when
+			 * a full MCV scan is not required. Otherwise, continue
+			 * scanning the remaining MCVs and accumulate the selectivity
+			 * of all matching MCVs.
+			 *
+			 * While scanning, also accumulate the selectivity of the
+			 * MCVs examined so far. This is used later to estimate the
+			 * selectivity of a non-NULL constant that does not match
+			 * any MCV.
+			 */
+			selec = 0.0;
 			for (i = 0; i < sslot.nvalues; i++)
 			{
 				Datum		fresult;
@@ -458,9 +487,21 @@ var_eq_const(VariableStatData *vardata, Oid oproid, Oid collation,
 				fresult = FunctionCallInvoke(fcinfo);
 				if (!fcinfo->isnull && DatumGetBool(fresult))
 				{
+					/*
+					 * Constant is "=" to this common value.  We know selectivity
+					 * exactly (or as exactly as ANALYZE could calculate it, anyway).
+					 */
 					match = true;
-					break;
+					if (!scan_entire_mcv)
+					{
+						selec = sslot.numbers[i];
+						break;
+					}
+
+					selec += sslot.numbers[i];
 				}
+
+				sumcommon += sslot.numbers[i];
 			}
 		}
 		else
@@ -469,26 +510,15 @@ var_eq_const(VariableStatData *vardata, Oid oproid, Oid collation,
 			i = 0;				/* keep compiler quiet */
 		}
 
-		if (match)
-		{
-			/*
-			 * Constant is "=" to this common value.  We know selectivity
-			 * exactly (or as exactly as ANALYZE could calculate it, anyway).
-			 */
-			selec = sslot.numbers[i];
-		}
-		else
+		if (!match)
 		{
 			/*
 			 * Comparison is against a constant that is neither NULL nor any
 			 * of the common values.  Its selectivity cannot be more than
 			 * this:
 			 */
-			double		sumcommon = 0.0;
 			double		otherdistinct;
 
-			for (i = 0; i < sslot.nnumbers; i++)
-				sumcommon += sslot.numbers[i];
 			selec = 1.0 - sumcommon - nullfrac;
 			CLAMP_PROBABILITY(selec);
 
-- 
2.43.0