expression: support builtin function json_contains (#7443)

expression/builtin.go

@@ -572,14 +572,15 @@ var funcs = map[string]functionClass{
 	ast.ValidatePasswordStrength: &validatePasswordStrengthFunctionClass{baseFunctionClass{ast.ValidatePasswordStrength, 1, 1}},
 
 	// json functions
-	ast.JSONType:    &jsonTypeFunctionClass{baseFunctionClass{ast.JSONType, 1, 1}},
-	ast.JSONExtract: &jsonExtractFunctionClass{baseFunctionClass{ast.JSONExtract, 2, -1}},
-	ast.JSONUnquote: &jsonUnquoteFunctionClass{baseFunctionClass{ast.JSONUnquote, 1, 1}},
-	ast.JSONSet:     &jsonSetFunctionClass{baseFunctionClass{ast.JSONSet, 3, -1}},
-	ast.JSONInsert:  &jsonInsertFunctionClass{baseFunctionClass{ast.JSONInsert, 3, -1}},
-	ast.JSONReplace: &jsonReplaceFunctionClass{baseFunctionClass{ast.JSONReplace, 3, -1}},
-	ast.JSONRemove:  &jsonRemoveFunctionClass{baseFunctionClass{ast.JSONRemove, 2, -1}},
-	ast.JSONMerge:   &jsonMergeFunctionClass{baseFunctionClass{ast.JSONMerge, 2, -1}},
-	ast.JSONObject:  &jsonObjectFunctionClass{baseFunctionClass{ast.JSONObject, 0, -1}},
-	ast.JSONArray:   &jsonArrayFunctionClass{baseFunctionClass{ast.JSONArray, 0, -1}},
+	ast.JSONType:     &jsonTypeFunctionClass{baseFunctionClass{ast.JSONType, 1, 1}},
+	ast.JSONExtract:  &jsonExtractFunctionClass{baseFunctionClass{ast.JSONExtract, 2, -1}},
+	ast.JSONUnquote:  &jsonUnquoteFunctionClass{baseFunctionClass{ast.JSONUnquote, 1, 1}},
+	ast.JSONSet:      &jsonSetFunctionClass{baseFunctionClass{ast.JSONSet, 3, -1}},
+	ast.JSONInsert:   &jsonInsertFunctionClass{baseFunctionClass{ast.JSONInsert, 3, -1}},
+	ast.JSONReplace:  &jsonReplaceFunctionClass{baseFunctionClass{ast.JSONReplace, 3, -1}},
+	ast.JSONRemove:   &jsonRemoveFunctionClass{baseFunctionClass{ast.JSONRemove, 2, -1}},
+	ast.JSONMerge:    &jsonMergeFunctionClass{baseFunctionClass{ast.JSONMerge, 2, -1}},
+	ast.JSONObject:   &jsonObjectFunctionClass{baseFunctionClass{ast.JSONObject, 0, -1}},
+	ast.JSONArray:    &jsonArrayFunctionClass{baseFunctionClass{ast.JSONArray, 0, -1}},
+	ast.JSONContains: &jsonContainsFunctionClass{baseFunctionClass{ast.JSONContains, 2, 3}},
 }

expression/builtin_json.go

@@ -35,6 +35,7 @@ var (
 	_ functionClass = &jsonMergeFunctionClass{}
 	_ functionClass = &jsonObjectFunctionClass{}
 	_ functionClass = &jsonArrayFunctionClass{}
+	_ functionClass = &jsonContainsFunctionClass{}
 
 	// Type of JSON value.
 	_ builtinFunc = &builtinJSONTypeSig{}
@@ -56,6 +57,8 @@ var (
 	_ builtinFunc = &builtinJSONRemoveSig{}
 	// Merge JSON documents, preserving duplicate keys.
 	_ builtinFunc = &builtinJSONMergeSig{}
+	// Check JSON document contains specific target.
+	_ builtinFunc = &builtinJSONContainsSig{}
 )
 
 type jsonTypeFunctionClass struct {
@@ -548,3 +551,66 @@ func jsonModify(ctx sessionctx.Context, args []Expression, row chunk.Row, mt jso
 	}
 	return res, false, nil
 }
+
+type jsonContainsFunctionClass struct {
+	baseFunctionClass
+}
+
+type builtinJSONContainsSig struct {
+	baseBuiltinFunc
+}
+
+func (b *builtinJSONContainsSig) Clone() builtinFunc {
+	newSig := &builtinJSONContainsSig{}
+	newSig.cloneFrom(&b.baseBuiltinFunc)
+	return newSig
+}
+
+func (c *jsonContainsFunctionClass) getFunction(ctx sessionctx.Context, args []Expression) (builtinFunc, error) {
+	if err := c.verifyArgs(args); err != nil {
+		return nil, errors.Trace(err)
+	}
+	argTps := []types.EvalType{types.ETJson, types.ETJson}
+	if len(args) == 3 {
+		argTps = append(argTps, types.ETString)
+	}
+	bf := newBaseBuiltinFuncWithTp(ctx, args, types.ETInt, argTps...)
+	sig := &builtinJSONContainsSig{bf}
+	sig.setPbCode(tipb.ScalarFuncSig_JsonContainsSig)
+	return sig, nil
+}
+
+func (b *builtinJSONContainsSig) evalInt(row chunk.Row) (res int64, isNull bool, err error) {
+	obj, isNull, err := b.args[0].EvalJSON(b.ctx, row)
+	if isNull || err != nil {
+		return res, isNull, errors.Trace(err)
+	}
+	target, isNull, err := b.args[1].EvalJSON(b.ctx, row)
+	if isNull || err != nil {
+		return res, isNull, errors.Trace(err)
+	}
+	var pathExpr json.PathExpression
+	if len(b.args) == 3 {
+		path, isNull, err := b.args[2].EvalString(b.ctx, row)
+		if isNull || err != nil {
+			return res, isNull, errors.Trace(err)
+		}
+		pathExpr, err = json.ParseJSONPathExpr(path)
+		if err != nil {
+			return res, true, errors.Trace(err)
+		}
+		if pathExpr.ContainsAnyAsterisk() {
+			return res, true, json.ErrInvalidJSONPathWildcard
+		}
+		var exists bool
+		obj, exists = obj.Extract([]json.PathExpression{pathExpr})
+		if !exists {
+			return res, true, nil
+		}
+	}
+
+	if json.ContainsBinary(obj, target) {
+		return 1, false, nil
+	}
+	return 0, false, nil
+}

expression/builtin_json_test.go

@@ -259,7 +259,7 @@ func (s *testEvaluatorSuite) TestJSONObject(c *C) {
 	}
 }
 
-func (s *testEvaluatorSuite) TestJSONORemove(c *C) {
+func (s *testEvaluatorSuite) TestJSONRemove(c *C) {
 	defer testleak.AfterTest(c)()
 	fc := funcs[ast.JSONRemove]
 	tbl := []struct {
@@ -308,3 +308,66 @@ func (s *testEvaluatorSuite) TestJSONORemove(c *C) {
 		}
 	}
 }
+
+func (s *testEvaluatorSuite) TestJSONContains(c *C) {
+	defer testleak.AfterTest(c)()
+	fc := funcs[ast.JSONContains]
+	tbl := []struct {
+		input    []interface{}
+		expected interface{}
+		success  bool
+	}{
+		// Tests nil arguments
+		{[]interface{}{nil, `1`, "$.c"}, nil, true},
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, nil, "$.a[3]"}, nil, true},
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, nil}, nil, true},
+		// Tests with path expression
+		{[]interface{}{`[1,2,[1,[5,[3]]]]`, `[1,3]`, "$[2]"}, 1, true},
+		{[]interface{}{`[1,2,[1,[5,{"a":[2,3]}]]]`, `[1,{"a":[3]}]`, "$[2]"}, 1, true},
+		{[]interface{}{`[{"a":1}]`, `{"a":1}`, "$"}, 1, true},
+		{[]interface{}{`[{"a":1,"b":2}]`, `{"a":1,"b":2}`, "$"}, 1, true},
+		{[]interface{}{`[{"a":{"a":1},"b":2}]`, `{"a":1}`, "$.a"}, 0, true},
+		// Tests without path expression
+		{[]interface{}{`{}`, `{}`}, 1, true},
+		{[]interface{}{`{"a":1}`, `{}`}, 1, true},
+		{[]interface{}{`{"a":1}`, `1`}, 0, true},
+		{[]interface{}{`{"a":[1]}`, `[1]`}, 0, true},
+		{[]interface{}{`{"b":2, "c":3}`, `{"c":3}`}, 1, true},
+		{[]interface{}{`1`, `1`}, 1, true},
+		{[]interface{}{`[1]`, `1`}, 1, true},
+		{[]interface{}{`[1,2]`, `[1]`}, 1, true},
+		{[]interface{}{`[1,2]`, `[1,3]`}, 0, true},
+		{[]interface{}{`[1,2]`, `["1"]`}, 0, true},
+		{[]interface{}{`[1,2,[1,3]]`, `[1,3]`}, 1, true},
+		{[]interface{}{`[1,2,[1,[5,[3]]]]`, `[1,3]`}, 1, true},
+		{[]interface{}{`[1,2,[1,[5,{"a":[2,3]}]]]`, `[1,{"a":[3]}]`}, 1, true},
+		{[]interface{}{`[{"a":1}]`, `{"a":1}`}, 1, true},
+		{[]interface{}{`[{"a":1,"b":2}]`, `{"a":1}`}, 1, true},
+		{[]interface{}{`[{"a":{"a":1},"b":2}]`, `{"a":1}`}, 0, true},
+		// Tests path expression contains any asterisk
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, "$.*"}, nil, false},
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, "$[*]"}, nil, false},
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, "$**.a"}, nil, false},
+		// Tests path expression does not identify a section of the target document
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, "$.c"}, nil, true},
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, "$.a[3]"}, nil, true},
+		{[]interface{}{`{"a": [1, 2, {"aa": "xx"}]}`, `1`, "$.a[2].b"}, nil, true},
+	}
+	for _, t := range tbl {
+		args := types.MakeDatums(t.input...)
+		f, err := fc.getFunction(s.ctx, s.datumsToConstants(args))
+		c.Assert(err, IsNil)
+		d, err := evalBuiltinFunc(f, chunk.Row{})
+
+		if t.success {
+			c.Assert(err, IsNil)
+			if t.expected == nil {
+				c.Assert(d.IsNull(), IsTrue)
+			} else {
+				c.Assert(d.GetInt64(), Equals, int64(t.expected.(int)))
+			}
+		} else {
+			c.Assert(err, NotNil)
+		}
+	}
+}

expression/distsql_builtin.go

@@ -430,6 +430,8 @@ func getSignatureByPB(ctx sessionctx.Context, sigCode tipb.ScalarFuncSig, tp *ti
 		f = &builtinJSONRemoveSig{base}
 	case tipb.ScalarFuncSig_JsonMergeSig:
 		f = &builtinJSONMergeSig{base}
+	case tipb.ScalarFuncSig_JsonContainsSig:
+		f = &builtinJSONContainsSig{base}
 	case tipb.ScalarFuncSig_LikeSig:
 		f = &builtinLikeSig{base}
 

expression/integration_test.go

@@ -3205,6 +3205,19 @@ func (s *testIntegrationSuite) TestFuncJSON(c *C) {
 	tk.MustExec(`update table_json set a=json_set(a,'$.a',json_object('a',1,'b',2)) where json_extract(a,'$.a[1]') = '2'`)
 	r = tk.MustQuery(`select json_extract(a, '$.a.a'), json_extract(a, '$.a.b') from table_json`)
 	r.Check(testkit.Rows("1 2", "<nil> <nil>"))
+
+	r = tk.MustQuery(`select json_contains(NULL, '1'), json_contains('1', NULL), json_contains('1', '1', NULL)`)
+	r.Check(testkit.Rows("<nil> <nil> <nil>"))
+	r = tk.MustQuery(`select json_contains('{}','{}'), json_contains('[1]','1'), json_contains('[1]','"1"'), json_contains('[1,2,[1,[5,[3]]]]', '[1,3]', '$[2]'), json_contains('[1,2,[1,[5,{"a":[2,3]}]]]', '[1,{"a":[3]}]', "$[2]"), json_contains('{"a":1}', '{"a":1,"b":2}', "$")`)
+	r.Check(testkit.Rows("1 1 0 1 1 0"))
+	r = tk.MustQuery(`select json_contains('{"a": 1}', '1', "$.c"), json_contains('{"a": [1, 2]}', '1', "$.a[2]"), json_contains('{"a": [1, {"a": 1}]}', '1', "$.a[1].b")`)
+	r.Check(testkit.Rows("<nil> <nil> <nil>"))
+	rs, err := tk.Exec("select json_contains('1','1','$.*')")
+	c.Assert(err, IsNil)
+	c.Assert(rs, NotNil)
+	_, err = session.GetRows4Test(context.Background(), tk.Se, rs)
+	c.Assert(err, NotNil)
+	c.Assert(err.Error(), Equals, "[json:3149]In this situation, path expressions may not contain the * and ** tokens.")
 }
 
 func (s *testIntegrationSuite) TestColumnInfoModified(c *C) {

mysql/errcode.go

@@ -889,6 +889,7 @@ const (
 	ErrInvalidJSONText                                              = 3140
 	ErrInvalidJSONPath                                              = 3143
 	ErrInvalidJSONData                                              = 3146
+	ErrInvalidJSONPathWildcard                                      = 3149
 	ErrJSONUsedAsKey                                                = 3152
 
 	// TiDB self-defined errors.

mysql/errname.go

@@ -886,6 +886,7 @@ var MySQLErrName = map[uint16]string{
 	ErrInvalidJSONText:                                       "Invalid JSON text: %-.192s",
 	ErrInvalidJSONPath:                                       "Invalid JSON path expression %s.",
 	ErrInvalidJSONData:                                       "Invalid data type for JSON data",
+	ErrInvalidJSONPathWildcard:                               "In this situation, path expressions may not contain the * and ** tokens.",
 	ErrJSONUsedAsKey:                                         "JSON column '%-.192s' cannot be used in key specification.",
 
 	// TiDB errors.

mysql/state.go

@@ -253,5 +253,6 @@ var MySQLState = map[uint16]string{
 	ErrInvalidJSONText:                     "22032",
 	ErrInvalidJSONPath:                     "42000",
 	ErrInvalidJSONData:                     "22032",
+	ErrInvalidJSONPathWildcard:             "42000",
 	ErrJSONUsedAsKey:                       "42000",
 }

types/json/binary_functions.go

@@ -689,3 +689,45 @@ func PeekBytesAsJSON(b []byte) (n int, err error) {
 	err = errors.New("Invalid JSON bytes")
 	return
 }
+
+// ContainsBinary check whether JSON document contains specific target according the following rules:
+// 1) object contains a target object if and only if every key is contained in source object and the value associated with the target key is contained in the value associated with the source key;
+// 2) array contains a target nonarray if and only if the target is contained in some element of the array;
+// 3) array contains a target array if and only if every element is contained in some element of the array;
+// 4) scalar contains a target scalar if and only if they are comparable and are equal;
+func ContainsBinary(obj, target BinaryJSON) bool {
+	switch obj.TypeCode {
+	case TypeCodeObject:
+		if target.TypeCode == TypeCodeObject {
+			len := target.getElemCount()
+			for i := 0; i < len; i++ {
+				key := target.objectGetKey(i)
+				val := target.objectGetVal(i)
+				if exp, exists := obj.objectSearchKey(key); !exists || !ContainsBinary(exp, val) {
+					return false
+				}
+			}
+			return true
+		}
+		return false
+	case TypeCodeArray:
+		if target.TypeCode == TypeCodeArray {
+			len := target.getElemCount()
+			for i := 0; i < len; i++ {
+				if !ContainsBinary(obj, target.arrayGetElem(i)) {
+					return false
+				}
+			}
+			return true
+		}
+		len := obj.getElemCount()
+		for i := 0; i < len; i++ {
+			if ContainsBinary(obj.arrayGetElem(i), target) {
+				return true
+			}
+		}
+		return false
+	default:
+		return CompareBinary(obj, target) == 0
+	}
+}

types/json/binary_test.go

@@ -293,3 +293,34 @@ func BenchmarkBinaryMarshal(b *testing.B) {
 		bj.MarshalJSON()
 	}
 }
+
+func (s *testJSONSuite) TestBinaryJSONContains(c *C) {
+	var tests = []struct {
+		input    string
+		target   string
+		expected bool
+	}{
+		{`{}`, `{}`, true},
+		{`{"a":1}`, `{}`, true},
+		{`{"a":1}`, `1`, false},
+		{`{"a":[1]}`, `[1]`, false},
+		{`{"b":2, "c":3}`, `{"c":3}`, true},
+		{`1`, `1`, true},
+		{`[1]`, `1`, true},
+		{`[1,2]`, `[1]`, true},
+		{`[1,2]`, `[1,3]`, false},
+		{`[1,2]`, `["1"]`, false},
+		{`[1,2,[1,3]]`, `[1,3]`, true},
+		{`[1,2,[1,[5,[3]]]]`, `[1,3]`, true},
+		{`[1,2,[1,[5,{"a":[2,3]}]]]`, `[1,{"a":[3]}]`, true},
+		{`[{"a":1}]`, `{"a":1}`, true},
+		{`[{"a":1,"b":2}]`, `{"a":1}`, true},
+		{`[{"a":{"a":1},"b":2}]`, `{"a":1}`, false},
+	}
+
+	for _, tt := range tests {
+		obj := mustParseBinaryFromString(c, tt.input)
+		target := mustParseBinaryFromString(c, tt.target)
+		c.Assert(ContainsBinary(obj, target), Equals, tt.expected)
+	}
+}

types/json/constants.go

@@ -212,12 +212,15 @@ var (
 	ErrInvalidJSONPath = terror.ClassJSON.New(mysql.ErrInvalidJSONPath, mysql.MySQLErrName[mysql.ErrInvalidJSONPath])
 	// ErrInvalidJSONData means invalid JSON data.
 	ErrInvalidJSONData = terror.ClassJSON.New(mysql.ErrInvalidJSONData, mysql.MySQLErrName[mysql.ErrInvalidJSONData])
+	// ErrInvalidJSONPathWildcard means invalid JSON path that contain wildcard characters.
+	ErrInvalidJSONPathWildcard = terror.ClassJSON.New(mysql.ErrInvalidJSONPathWildcard, mysql.MySQLErrName[mysql.ErrInvalidJSONPathWildcard])
 )
 
 func init() {
 	terror.ErrClassToMySQLCodes[terror.ClassJSON] = map[terror.ErrCode]uint16{
-		mysql.ErrInvalidJSONText: mysql.ErrInvalidJSONText,
-		mysql.ErrInvalidJSONPath: mysql.ErrInvalidJSONPath,
-		mysql.ErrInvalidJSONData: mysql.ErrInvalidJSONData,
+		mysql.ErrInvalidJSONText:         mysql.ErrInvalidJSONText,
+		mysql.ErrInvalidJSONPath:         mysql.ErrInvalidJSONPath,
+		mysql.ErrInvalidJSONData:         mysql.ErrInvalidJSONData,
+		mysql.ErrInvalidJSONPathWildcard: mysql.ErrInvalidJSONPathWildcard,
 	}
 }

types/json/path_expr.go

@@ -117,6 +117,11 @@ func (pe PathExpression) popOneLastLeg() (PathExpression, pathLeg) {
 	return PathExpression{legs: pe.legs[:lastLegIdx]}, lastLeg
 }
 
+// ContainsAnyAsterisk returns true if pe contains any asterisk.
+func (pe PathExpression) ContainsAnyAsterisk() bool {
+	return pe.flags.containsAnyAsterisk()
+}
+
 // ParseJSONPathExpr parses a JSON path expression. Returns a PathExpression
 // object which can be used in JSON_EXTRACT, JSON_SET and so on.
 func ParseJSONPathExpr(pathExpr string) (pe PathExpression, err error) {