expression: let cast function supports explicit set charset (#55724)

expression/bench_test.go

@@ -1452,7 +1452,7 @@ func genVecBuiltinFuncBenchCase(ctx sessionctx.Context, funcName string, testCas
 		case types.ETJson:
 			fc = &castAsJSONFunctionClass{baseFunctionClass{ast.Cast, 1, 1}, tp}
 		case types.ETString:
-			fc = &castAsStringFunctionClass{baseFunctionClass{ast.Cast, 1, 1}, tp}
+			fc = &castAsStringFunctionClass{baseFunctionClass{ast.Cast, 1, 1}, tp, false}
 		}
 		baseFunc, err = fc.getFunction(ctx, cols)
 	} else if funcName == ast.GetVar {

expression/builtin.go

@@ -418,6 +418,36 @@
 	}
 }
 
+func newBaseBuiltinCastFunc4String(ctx sessionctx.Context, funcName string, args []Expression, tp *types.FieldType, isExplicitCharset bool) (baseBuiltinFunc, error) {
+	var bf baseBuiltinFunc
+	var err error
+	if isExplicitCharset {
+		bf = baseBuiltinFunc{
+			bufAllocator:           newLocalColumnPool(),
+			childrenVectorizedOnce: new(sync.Once),
+
+			args: args,
+			ctx:  ctx,
+			tp:   tp,
+		}
+		bf.SetCharsetAndCollation(tp.GetCharset(), tp.GetCollate())
+		bf.setCollator(collate.GetCollator(tp.GetCollate()))
+		bf.SetCoercibility(CoercibilityExplicit)
+		bf.SetExplicitCharset(true)
+		if tp.GetCharset() == charset.CharsetASCII {
+			bf.SetRepertoire(ASCII)
+		} else {
+			bf.SetRepertoire(UNICODE)
+		}
+	} else {
+		bf, err = newBaseBuiltinFunc(ctx, funcName, args, tp)
+		if err != nil {
+			return baseBuiltinFunc{}, err
+		}
+	}
+	return bf, nil
+}
+
 // vecBuiltinFunc contains all vectorized methods for a builtin function.
 type vecBuiltinFunc interface {
 	// vectorized returns if this builtin function itself supports vectorized evaluation.

expression/builtin_cast.go

@@ -272,14 +272,15 @@ func (c *castAsDecimalFunctionClass) getFunction(ctx sessionctx.Context, args []
 type castAsStringFunctionClass struct {
 	baseFunctionClass
 
-	tp *types.FieldType
+	tp                *types.FieldType
+	isExplicitCharset bool
 }
 
 func (c *castAsStringFunctionClass) getFunction(ctx sessionctx.Context, args []Expression) (sig builtinFunc, err error) {
 	if err := c.verifyArgs(args); err != nil {
 		return nil, err
 	}
-	bf, err := newBaseBuiltinFunc(ctx, c.funcName, args, c.tp)
+	bf, err := newBaseBuiltinCastFunc4String(ctx, c.funcName, args, c.tp, c.isExplicitCharset)
 	if err != nil {
 		return nil, err
 	}
@@ -2080,13 +2081,13 @@ func BuildCastCollationFunction(ctx sessionctx.Context, expr Expression, ec *Exp
 
 // BuildCastFunction builds a CAST ScalarFunction from the Expression.
 func BuildCastFunction(ctx sessionctx.Context, expr Expression, tp *types.FieldType) (res Expression) {
-	res, err := BuildCastFunctionWithCheck(ctx, expr, tp)
+	res, err := BuildCastFunctionWithCheck(ctx, expr, tp, false)
 	terror.Log(err)
 	return
 }
 
 // BuildCastFunctionWithCheck builds a CAST ScalarFunction from the Expression and return error if any.
-func BuildCastFunctionWithCheck(ctx sessionctx.Context, expr Expression, tp *types.FieldType) (res Expression, err error) {
+func BuildCastFunctionWithCheck(ctx sessionctx.Context, expr Expression, tp *types.FieldType, isExplicitCharset bool) (res Expression, err error) {
 	argType := expr.GetType()
 	// If source argument's nullable, then target type should be nullable
 	if !mysql.HasNotNullFlag(argType.GetFlag()) {
@@ -2112,7 +2113,7 @@ func BuildCastFunctionWithCheck(ctx sessionctx.Context, expr Expression, tp *typ
 			fc = &castAsJSONFunctionClass{baseFunctionClass{ast.Cast, 1, 1}, tp}
 		}
 	case types.ETString:
-		fc = &castAsStringFunctionClass{baseFunctionClass{ast.Cast, 1, 1}, tp}
+		fc = &castAsStringFunctionClass{baseFunctionClass{ast.Cast, 1, 1}, tp, isExplicitCharset}
 		if expr.GetType().GetType() == mysql.TypeBit {
 			tp.SetFlen((expr.GetType().GetFlen() + 7) / 8)
 		}

expression/builtin_cast_test.go

@@ -646,7 +646,7 @@ func TestCastFuncSig(t *testing.T) {
 		tp := types.NewFieldType(mysql.TypeVarString)
 		tp.SetCharset(charset.CharsetBin)
 		args := []Expression{c.before}
-		stringFunc, err := newBaseBuiltinFunc(ctx, "", args, tp)
+		stringFunc, err := newBaseBuiltinCastFunc4String(ctx, "", args, tp, false)
 		require.NoError(t, err)
 		switch i {
 		case 0:
@@ -732,7 +732,7 @@ func TestCastFuncSig(t *testing.T) {
 		tp := types.NewFieldType(mysql.TypeVarString)
 		tp.SetFlen(c.flen)
 		tp.SetCharset(charset.CharsetBin)
-		stringFunc, err := newBaseBuiltinFunc(ctx, "", args, tp)
+		stringFunc, err := newBaseBuiltinCastFunc4String(ctx, "", args, tp, false)
 		require.NoError(t, err)
 		switch i {
 		case 0:
@@ -1083,7 +1083,7 @@ func TestCastFuncSig(t *testing.T) {
 	// null case
 	args := []Expression{&Column{RetType: types.NewFieldType(mysql.TypeDouble), Index: 0}}
 	row := chunk.MutRowFromDatums([]types.Datum{types.NewDatum(nil)})
-	bf, err := newBaseBuiltinFunc(ctx, "", args, types.NewFieldType(mysql.TypeVarString))
+	bf, err := newBaseBuiltinCastFunc4String(ctx, "", args, types.NewFieldType(mysql.TypeVarString), false)
 	require.NoError(t, err)
 	sig = &builtinCastRealAsStringSig{bf}
 	sRes, isNull, err := sig.evalString(row.ToRow())
@@ -1677,10 +1677,8 @@ func TestCastArrayFunc(t *testing.T) {
 		},
 	}
 	for _, tt := range tbl {
-		f, err := BuildCastFunctionWithCheck(ctx, datumsToConstants(types.MakeDatums(types.CreateBinaryJSON(tt.input)))[0], tt.tp)
-		if tt.buildFuncSuccess {
-			require.NoError(t, err, tt.input)
-		} else {
+		f, err := BuildCastFunctionWithCheck(ctx, datumsToConstants(types.MakeDatums(types.CreateBinaryJSON(tt.input)))[0], tt.tp, false)
+		if !tt.buildFuncSuccess {
 			require.Error(t, err, tt.input)
 			continue
 		}

expression/collation.go

@@ -44,6 +44,8 @@ type collationInfo struct {
 
 	charset   string
 	collation string
+
+	isExplicitCharset bool
 }
 
 func (c *collationInfo) HasCoercibility() bool {
@@ -76,6 +78,14 @@ func (c *collationInfo) CharsetAndCollation() (string, string) {
 	return c.charset, c.collation
 }
 
+func (c *collationInfo) IsExplicitCharset() bool {
+	return c.isExplicitCharset
+}
+
+func (c *collationInfo) SetExplicitCharset(explicit bool) {
+	c.isExplicitCharset = explicit
+}
+
 // CollationInfo contains all interfaces about dealing with collation.
 type CollationInfo interface {
 	// HasCoercibility returns if the Coercibility value is initialized.
@@ -98,6 +108,12 @@ type CollationInfo interface {
 
 	// SetCharsetAndCollation sets charset and collation.
 	SetCharsetAndCollation(chs, coll string)
+
+	// IsExplicitCharset return the charset is explicit set or not.
+	IsExplicitCharset() bool
+
+	// SetExplicitCharset set the charset is explicit or not.
+	SetExplicitCharset(bool)
 }
 
 // Coercibility values are used to check whether the collation of one item can be coerced to
@@ -246,9 +262,8 @@ func deriveCollation(ctx sessionctx.Context, funcName string, args []Expression,
 	case ast.Cast:
 		// We assume all the cast are implicit.
 		ec = &ExprCollation{args[0].Coercibility(), args[0].Repertoire(), args[0].GetType().GetCharset(), args[0].GetType().GetCollate()}
-		// Non-string type cast to string type should use @@character_set_connection and @@collation_connection.
-		// String type cast to string type should keep its original charset and collation. It should not happen.
-		if retType == types.ETString && argTps[0] != types.ETString {
+		// Cast to string type should use @@character_set_connection and @@collation_connection.
+		if retType == types.ETString {
 			ec.Charset, ec.Collation = ctx.GetSessionVars().GetCharsetInfo()
 		}
 		return ec, nil

expression/scalar_function.go

@@ -626,6 +626,16 @@
 	sf.Function.SetRepertoire(r)
 }
 
+// IsExplicitCharset return the charset is explicit set or not.
+func (sf *ScalarFunction) IsExplicitCharset() bool {
+	return sf.Function.IsExplicitCharset()
+}
+
+// SetExplicitCharset set the charset is explicit or not.
+func (sf *ScalarFunction) SetExplicitCharset(explicit bool) {
+	sf.Function.SetExplicitCharset(explicit)
+}
+
 const emptyScalarFunctionSize = int64(unsafe.Sizeof(ScalarFunction{}))
 
 // MemoryUsage return the memory usage of ScalarFunction

expression/util.go

@@ -453,7 +453,8 @@ func ColumnSubstituteImpl(expr Expression, schema *Schema, newExprs []Expression
 			}
 			if substituted {
 				flag := v.RetType.GetFlag()
-				e := BuildCastFunction(v.GetCtx(), newArg, v.RetType)
+				e, err := BuildCastFunctionWithCheck(v.GetCtx(), newArg, v.RetType, v.Function.IsExplicitCharset())
+				terror.Log(err)
 				e.SetCoercibility(v.Coercibility())
 				e.GetType().SetFlag(flag)
 				return true, false, e

expression/util_test.go

@@ -588,6 +588,8 @@ func (m *MockExpr) Coercibility() Coercibility
 func (m *MockExpr) SetCoercibility(Coercibility)                                  {}
 func (m *MockExpr) Repertoire() Repertoire                                        { return UNICODE }
 func (m *MockExpr) SetRepertoire(Repertoire)                                      {}
+func (m *MockExpr) IsExplicitCharset() bool                                       { return false }
+func (m *MockExpr) SetExplicitCharset(bool)                                       {}
 
 func (m *MockExpr) CharsetAndCollation() (string, string) {
 	return "", ""

planner/core/expression_rewriter.go

@@ -1209,7 +1209,7 @@ func (er *expressionRewriter) Leave(originInNode ast.Node) (retNode ast.Node, ok
 			return retNode, false
 		}
 
-		castFunction, err := expression.BuildCastFunctionWithCheck(er.sctx, arg, v.Tp)
+		castFunction, err := expression.BuildCastFunctionWithCheck(er.sctx, arg, v.Tp, v.ExplicitCharSet)
 		if err != nil {
 			er.err = err
 			return retNode, false

tests/integrationtest/r/expression/cast.result

@@ -28,3 +28,33 @@ select 1.194192591e9 > t0.c0 from t0;
 select 1.194192591e9 < t0.c0 from t0;
 1.194192591e9 < t0.c0
 0
+drop table if exists test;
+CREATE TABLE `test` (
+`id` bigint(20) NOT NULL,
+`update_user` varchar(32) DEFAULT NULL,
+PRIMARY KEY (`id`) /*T![clustered_index] CLUSTERED */
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin;
+insert into test values(1,'张三');
+insert into test values(2,'李四');
+insert into test values(3,'张三');
+insert into test values(4,'李四');
+select * from test order by cast(update_user as char character set gbk) desc , id limit 3;
+id	update_user
+1	张三
+3	张三
+2	李四
+drop table test;
+CREATE TABLE `test` (
+`id` bigint NOT NULL,
+`update_user` varchar(32) CHARACTER SET gbk COLLATE gbk_chinese_ci DEFAULT NULL,
+PRIMARY KEY (`id`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin;
+insert into test values(1,'张三');
+insert into test values(2,'李四');
+insert into test values(3,'张三');
+insert into test values(4,'李四');
+select * from test order by cast(update_user as char) desc , id limit 3;
+id	update_user
+2	李四
+4	李四
+1	张三

tests/integrationtest/t/expression/cast.test

@@ -11,3 +11,29 @@ select t0.c0 > 1.194192591e9 from t0;
 select t0.c0 < 1.194192591e9 from t0;
 select 1.194192591e9 > t0.c0 from t0;
 select 1.194192591e9 < t0.c0 from t0;
+
+# TestCastAsStringExplicitCharSet
+drop table if exists test;
+CREATE TABLE `test` (
+    `id` bigint(20) NOT NULL,
+    `update_user` varchar(32) DEFAULT NULL,
+    PRIMARY KEY (`id`) /*T![clustered_index] CLUSTERED */
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin;
+insert into test values(1,'张三');
+insert into test values(2,'李四');
+insert into test values(3,'张三');
+insert into test values(4,'李四');
+select * from test order by cast(update_user as char character set gbk) desc , id limit 3;
+
+drop table test;
+CREATE TABLE `test` (
+  `id` bigint NOT NULL,
+  `update_user` varchar(32) CHARACTER SET gbk COLLATE gbk_chinese_ci DEFAULT NULL,
+  PRIMARY KEY (`id`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin;
+insert into test values(1,'张三');
+insert into test values(2,'李四');
+insert into test values(3,'张三');
+insert into test values(4,'李四');
+select * from test order by cast(update_user as char) desc , id limit 3;
+