planner: address collation ambiguity in scalar function construction during predicate simplification. (#57049) (#57476)

close #56479

Author: ti-chi-bot

pkg/planner/core/rule_predicate_simplification.go

@@ -221,7 +221,10 @@ func unsatisfiable(ctx base.PlanContext, p1, p2 expression.Expression) bool {
 	// Copy constant from equal predicate into other predicate.
 	equalValue := equalPred.(*expression.ScalarFunction)
 	otherValue := otherPred.(*expression.ScalarFunction)
-	newPred := expression.NewFunctionInternal(ctx.GetExprCtx(), otherValue.FuncName.L, otherValue.RetType, equalValue.GetArgs()[1], otherValue.GetArgs()[1])
+	newPred, err := expression.NewFunction(ctx.GetExprCtx(), otherValue.FuncName.L, otherValue.RetType, equalValue.GetArgs()[1], otherValue.GetArgs()[1])
+	if err != nil {
+		return false
+	}
 	newPredList := make([]expression.Expression, 0, 1)
 	newPredList = append(newPredList, newPred)
 	newPredList = expression.PropagateConstant(ctx.GetExprCtx(), newPredList)
@@ -273,7 +276,11 @@ func updateOrPredicate(ctx base.PlanContext, orPredicateList expression.Expressi
 	} else if emptyFirst && emptySecond {
 		return &expression.Constant{Value: types.NewIntDatum(0), RetType: types.NewFieldType(mysql.TypeTiny)}
 	}
-	return expression.NewFunctionInternal(ctx.GetExprCtx(), ast.LogicOr, v.RetType, firstCondition, secondCondition)
+	newPred, err := expression.NewFunction(ctx.GetExprCtx(), ast.LogicOr, v.RetType, firstCondition, secondCondition)
+	if err != nil {
+		return orPredicateList
+	}
+	return newPred
 }
 
 // pruneEmptyORBranches applies iteratively updateOrPredicate for each pair of OR predicate

tests/integrationtest/r/planner/core/issuetest/planner_issue.result

@@ -750,3 +750,12 @@ NULL	NULL	2	2	4	2
 show warnings;
 Level	Code	Message
 drop table if exists t1, t2, t3, t4;
+drop table if exists tl6e913fb9;
+CREATE TABLE `tl6e913fb9` (
+`col_36` varchar(175) COLLATE utf8mb4_unicode_ci NOT NULL DEFAULT 'asMF',
+KEY `idx_35_5` (`col_36`(1)),
+PRIMARY KEY (`col_36`) /*T![clustered_index] NONCLUSTERED */,
+KEY `idx_65` (`col_36`(5))
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
+with cte_192 ( col_1101,col_1102,col_1103,col_1104 ) AS ( select  /*+ use_index_merge( tl6e913fb9 ) */   replace( tl6e913fb9.col_36 , tl6e913fb9.col_36 , tl6e913fb9.col_36 ) as r0 , space( 0 ) as r1 , min( distinct  tl6e913fb9.col_36 ) as r2 , count( distinct  tl6e913fb9.col_36 ) as r3 from tl6e913fb9 where tl6e913fb9.col_36 between 'n92ok$B%W#UU%O' and '()c=KVQ=T%-vzGJ' and tl6e913fb9.col_36 in ( 'T+kf' ,'Lvluod2H' ,'3#Omx@pC^fFkeH' ,'=b$z' ) group by tl6e913fb9.col_36  having tl6e913fb9.col_36 = 'xjV@' or IsNull( tl6e913fb9.col_36 ) ) ( select 1,col_1101,col_1102,col_1103,col_1104 from cte_192 where not( IsNull( cte_192.col_1102 ) ) order by 1,2,3,4,5 limit 72850972 );
+1	col_1101	col_1102	col_1103	col_1104

tests/integrationtest/t/planner/core/issuetest/planner_issue.test

@@ -517,3 +517,13 @@ SELECT t2.a,t2.b,t3.a,t3.b,t4.a,t4.b
        ON t3.a=1 AND t3.b=t2.b AND t2.b=t4.b order by 1, 2, 3, 4, 5;
 show warnings;
 drop table if exists t1, t2, t3, t4;
+
+# TestIssue56479
+drop table if exists tl6e913fb9;
+CREATE TABLE `tl6e913fb9` (
+  `col_36` varchar(175) COLLATE utf8mb4_unicode_ci NOT NULL DEFAULT 'asMF',
+  KEY `idx_35_5` (`col_36`(1)),
+  PRIMARY KEY (`col_36`) /*T![clustered_index] NONCLUSTERED */,
+  KEY `idx_65` (`col_36`(5))
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
+with cte_192 ( col_1101,col_1102,col_1103,col_1104 ) AS ( select  /*+ use_index_merge( tl6e913fb9 ) */   replace( tl6e913fb9.col_36 , tl6e913fb9.col_36 , tl6e913fb9.col_36 ) as r0 , space( 0 ) as r1 , min( distinct  tl6e913fb9.col_36 ) as r2 , count( distinct  tl6e913fb9.col_36 ) as r3 from tl6e913fb9 where tl6e913fb9.col_36 between 'n92ok$B%W#UU%O' and '()c=KVQ=T%-vzGJ' and tl6e913fb9.col_36 in ( 'T+kf' ,'Lvluod2H' ,'3#Omx@pC^fFkeH' ,'=b$z' ) group by tl6e913fb9.col_36  having tl6e913fb9.col_36 = 'xjV@' or IsNull( tl6e913fb9.col_36 ) ) ( select 1,col_1101,col_1102,col_1103,col_1104 from cte_192 where not( IsNull( cte_192.col_1102 ) ) order by 1,2,3,4,5 limit 72850972 );