planner: fix expression rewrite makes between expr infers wrong collation. (#27254)

expression/builtin.go

@@ -131,7 +131,7 @@ func illegalMixCollationErr(funcName string, args []Expression) error {
 	case 2:
 		return collate.ErrIllegalMix2Collation.GenWithStackByArgs(args[0].GetType().Collate, coerString[args[0].Coercibility()], args[1].GetType().Collate, coerString[args[1].Coercibility()], funcName)
 	case 3:
-		return collate.ErrIllegalMix3Collation.GenWithStackByArgs(args[0].GetType().Collate, coerString[args[0].Coercibility()], args[1].GetType().Collate, coerString[args[1].Coercibility()], args[0].GetType().Collate, coerString[args[2].Coercibility()], funcName)
+		return collate.ErrIllegalMix3Collation.GenWithStackByArgs(args[0].GetType().Collate, coerString[args[0].Coercibility()], args[1].GetType().Collate, coerString[args[1].Coercibility()], args[2].GetType().Collate, coerString[args[2].Coercibility()], funcName)
 	default:
 		return collate.ErrIllegalMixCollation.GenWithStackByArgs(funcName)
 	}

expression/integration_test.go

@@ -8662,7 +8662,7 @@ func (s *testIntegrationSerialSuite) TestLikeWithCollation(c *C) {
 	defer collate.SetNewCollationEnabledForTest(false)
 
 	tk.MustQuery(`select 'a' like 'A' collate utf8mb4_unicode_ci;`).Check(testkit.Rows("1"))
-	tk.MustGetErrMsg(`select 'a' collate utf8mb4_bin like 'A' collate utf8mb4_unicode_ci;`, "[expression:1270]Illegal mix of collations (utf8mb4_bin,EXPLICIT), (utf8mb4_unicode_ci,EXPLICIT), (utf8mb4_bin,NUMERIC) for operation 'like'")
+	tk.MustGetErrMsg(`select 'a' collate utf8mb4_bin like 'A' collate utf8mb4_unicode_ci;`, "[expression:1270]Illegal mix of collations (utf8mb4_bin,EXPLICIT), (utf8mb4_unicode_ci,EXPLICIT), (binary,NUMERIC) for operation 'like'")
 	tk.MustQuery(`select '😛' collate utf8mb4_general_ci like '😋';`).Check(testkit.Rows("1"))
 	tk.MustQuery(`select '😛' collate utf8mb4_general_ci = '😋';`).Check(testkit.Rows("1"))
 	tk.MustQuery(`select '😛' collate utf8mb4_unicode_ci like '😋';`).Check(testkit.Rows("0"))

planner/core/expression_rewriter.go

@@ -1596,17 +1596,25 @@ func (er *expressionRewriter) betweenToExpression(v *ast.BetweenExpr) {
 
 	expr, lexp, rexp := er.wrapExpWithCast()
 
-	var op string
+	er.err = expression.CheckIllegalMixCollation("between", []expression.Expression{expr, lexp, rexp}, types.ETInt)
+	if er.err != nil {
+		return
+	}
+
+	dstCharset, dstCollation := expression.DeriveCollationFromExprs(er.sctx, expr, lexp, rexp)
+
 	var l, r expression.Expression
-	l, er.err = er.newFunction(ast.GE, &v.Type, expr, lexp)
-	if er.err == nil {
-		r, er.err = er.newFunction(ast.LE, &v.Type, expr, rexp)
+	l, er.err = expression.NewFunctionBase(er.sctx, ast.GE, &v.Type, expr, lexp)
+	if er.err != nil {
+		return
 	}
-	op = ast.LogicAnd
+	r, er.err = expression.NewFunctionBase(er.sctx, ast.LE, &v.Type, expr, rexp)
 	if er.err != nil {
 		return
 	}
-	function, err := er.newFunction(op, &v.Type, l, r)
+	l.SetCharsetAndCollation(dstCharset, dstCollation)
+	r.SetCharsetAndCollation(dstCharset, dstCollation)
+	function, err := er.newFunction(ast.LogicAnd, &v.Type, l, r)
 	if err != nil {
 		er.err = err
 		return

planner/core/expression_rewriter_test.go

@@ -17,16 +17,21 @@ import (
 	. "github.com/pingcap/check"
 	"github.com/pingcap/parser/terror"
 	"github.com/pingcap/tidb/planner/core"
+	"github.com/pingcap/tidb/util/collate"
 	"github.com/pingcap/tidb/util/testkit"
 	"github.com/pingcap/tidb/util/testleak"
 	"github.com/pingcap/tidb/util/testutil"
 )
 
 var _ = Suite(&testExpressionRewriterSuite{})
+var _ = SerialSuites(&testExpressionRewriterSuiteSerial{})
 
 type testExpressionRewriterSuite struct {
 }
 
+type testExpressionRewriterSuiteSerial struct {
+}
+
 func (s *testExpressionRewriterSuite) TestIfNullEliminateColName(c *C) {
 	defer testleak.AfterTest(c)()
 	store, dom, err := newStoreWithBootstrap()
@@ -409,3 +414,82 @@ func (s *testExpressionRewriterSuite) TestIssue22818(c *C) {
 	tk.MustQuery("select * from t where a between \"23:22:22\" and \"23:22:22\"").Check(
 		testkit.Rows("23:22:22"))
 }
+<<<<<<< HEAD
+=======
+
+func (s *testExpressionRewriterSuite) TestIssue24705(c *C) {
+	defer testleak.AfterTest(c)()
+	store, dom, err := newStoreWithBootstrap()
+	c.Assert(err, IsNil)
+	tk := testkit.NewTestKit(c, store)
+	defer func() {
+		dom.Close()
+		store.Close()
+	}()
+
+	tk.MustExec("use test;")
+	tk.MustExec("drop table if exists t1,t2;")
+	tk.MustExec("create table t1 (c_int int, c_str varchar(40) character set utf8 collate utf8_general_ci);")
+	tk.MustExec("create table t2 (c_int int, c_str varchar(40) character set utf8 collate utf8_unicode_ci);")
+	err = tk.ExecToErr("select * from t1 where c_str < any (select c_str from t2 where c_int between 6 and 9);")
+	c.Assert(err.Error(), Equals, "[expression:1267]Illegal mix of collations (utf8_general_ci,IMPLICIT) and (utf8_unicode_ci,IMPLICIT) for operation '<'")
+}
+
+func (s *testExpressionRewriterSuiteSerial) TestBetweenExprCollation(c *C) {
+	collate.SetNewCollationEnabledForTest(true)
+	defer collate.SetNewCollationEnabledForTest(false)
+
+	defer testleak.AfterTest(c)()
+	store, dom, err := newStoreWithBootstrap()
+	c.Assert(err, IsNil)
+	tk := testkit.NewTestKit(c, store)
+	defer func() {
+		dom.Close()
+		store.Close()
+	}()
+
+	tk.MustExec("use test")
+	tk.MustExec("drop table if exists t1;")
+	tk.MustExec("create table t1(a char(10) charset latin1 collate latin1_bin, c char(10) collate utf8mb4_general_ci);")
+	tk.MustExec("insert into t1 values ('a', 'B');")
+	tk.MustExec("insert into t1 values ('c', 'D');")
+	tk.MustQuery("select * from t1 where a between 'B' and c;").Check(testkit.Rows("c D"))
+
+	tk.MustGetErrMsg("select * from t1 where a between 'B' collate utf8mb4_general_ci and c collate utf8mb4_unicode_ci;", "[expression:1270]Illegal mix of collations (latin1_bin,IMPLICIT), (utf8mb4_general_ci,EXPLICIT), (utf8mb4_unicode_ci,EXPLICIT) for operation 'between'")
+}
+
+func (s *testExpressionRewriterSuite) TestMultiColInExpression(c *C) {
+	store, dom, err := newStoreWithBootstrap()
+	c.Assert(err, IsNil)
+	tk := testkit.NewTestKit(c, store)
+	defer func() {
+		dom.Close()
+		store.Close()
+	}()
+
+	tk.MustExec("use test;")
+	tk.MustExec("drop table if exists t1, t2")
+	tk.MustExec("create table t1(a int, b int)")
+	tk.MustExec("insert into t1 values(1,1),(2,null),(null,3),(4,4)")
+	tk.MustExec("analyze table t1")
+	tk.MustExec("create table t2(a int, b int)")
+	tk.MustExec("insert into t2 values(1,1),(2,null),(null,3),(5,4)")
+	tk.MustExec("analyze table t2")
+	var input []string
+	var output []struct {
+		SQL  string
+		Plan []string
+		Res  []string
+	}
+	s.testData.GetTestCases(c, &input, &output)
+	for i, tt := range input {
+		s.testData.OnRecord(func() {
+			output[i].SQL = tt
+			output[i].Plan = s.testData.ConvertRowsToStrings(tk.MustQuery("explain format = 'brief' " + tt).Rows())
+			output[i].Res = s.testData.ConvertRowsToStrings(tk.MustQuery(tt).Sort().Rows())
+		})
+		tk.MustQuery("explain format = 'brief' " + tt).Check(testkit.Rows(output[i].Plan...))
+		tk.MustQuery(tt).Sort().Check(testkit.Rows(output[i].Res...))
+	}
+}
+>>>>>>> d5baa5dad... planner: fix expression rewrite makes between expr infers wrong collation. (#27254)