[fix](json) Add . after in JSON path to support correct token parsing (#52543) (#52544)

Boost tokenizer requires explicit "." after "$" to correctly extract JSON path tokens. Without this, expressions like "$[0].key" cannot be properly split, causing issues in downstream logic. This commit ensures a "." is automatically added after "$" to maintain consistent token parsing behavior.
2025-07-03 14:36:53 +08:00
parent fb70742e87
commit 6404277795
9 changed files with 428 additions and 440 deletions
--- a/be/src/vec/functions/function_json.cpp
+++ b/be/src/vec/functions/function_json.cpp
@ -145,45 +145,7 @@ rapidjson::Value* match_value(const std::vector<JsonPath>& parsed_paths, rapidjs
        const std::string& col = parsed_paths[i].key;
        int index = parsed_paths[i].idx;
        if (LIKELY(!col.empty())) {
-            if (root->IsArray()) {
-                array_obj = static_cast<rapidjson::Value*>(
-                        mem_allocator.Malloc(sizeof(rapidjson::Value)));
-                array_obj->SetArray();
-                bool is_null = true;
-
-                // if array ,loop the array,find out all Objects,then find the results from the objects
-                for (int j = 0; j < root->Size(); j++) {
-                    rapidjson::Value* json_elem = &((*root)[j]);
-
-                    if (json_elem->IsArray() || json_elem->IsNull()) {
-                        continue;
-                    } else {
-                        if (!json_elem->IsObject()) {
-                            continue;
-                        }
-                        if (!json_elem->HasMember(col.c_str())) {
-                            if (is_insert_null) { // not found item, then insert a null object.
-                                is_null = false;
-                                rapidjson::Value nullObject(rapidjson::kNullType);
-                                array_obj->PushBack(nullObject, mem_allocator);
-                            }
-                            continue;
-                        }
-                        rapidjson::Value* obj = &((*json_elem)[col.c_str()]);
-                        if (obj->IsArray()) {
-                            is_null = false;
-                            for (int k = 0; k < obj->Size(); k++) {
-                                array_obj->PushBack((*obj)[k], mem_allocator);
-                            }
-                        } else if (!obj->IsNull()) {
-                            is_null = false;
-                            array_obj->PushBack(*obj, mem_allocator);
-                        }
-                    }
-                }
-
-                root = is_null ? &(array_obj->SetNull()) : array_obj;
-            } else if (root->IsObject()) {
+            if (root->IsObject()) {
                if (!root->HasMember(col.c_str())) {
                    return nullptr;
                } else {
@ -234,8 +196,17 @@ rapidjson::Value* get_json_object(std::string_view json_string, std::string_view

    //Cannot use '\' as the last character, return NULL
    if (path_string.back() == '\\') {
-        document->SetNull();
-        return document;
+        return nullptr;
+    }
+
+    std::string fixed_string;
+    if (path_string.size() >= 2 && path_string[0] == '$' && path_string[1] != '.') {
+        // Boost tokenizer requires explicit "." after "$" to correctly extract JSON path tokens.
+        // Without this, expressions like "$[0].key" cannot be properly split.
+        // This commit ensures a "." is automatically added after "$" to maintain consistent token parsing behavior.
+        fixed_string = "$.";
+        fixed_string += path_string.substr(1);
+        path_string = fixed_string;
    }

    try {
@ -252,13 +223,13 @@ rapidjson::Value* get_json_object(std::string_view json_string, std::string_view
        }
    } catch (boost::escaped_list_error&) {
        // meet unknown escape sequence, example '$.name\k'
-        return document;
+        return nullptr;
    }

    parsed_paths = &tmp_parsed_paths;

    if (!(*parsed_paths)[0].is_valid) {
-        return document;
+        return nullptr;
    }

    if (UNLIKELY((*parsed_paths).size() == 1)) {
@ -271,10 +242,7 @@ rapidjson::Value* get_json_object(std::string_view json_string, std::string_view

    document->Parse(json_string.data(), json_string.size());
    if (UNLIKELY(document->HasParseError())) {
-        // VLOG_CRITICAL << "Error at offset " << document->GetErrorOffset() << ": "
-        //         << GetParseError_En(document->GetParseError());
-        document->SetNull();
-        return document;
+        return nullptr;
    }

    return match_value(*parsed_paths, document, document->GetAllocator());
@ -849,9 +817,9 @@ struct FunctionJsonQuoteImpl {
 struct FunctionJsonExtractImpl {
    static constexpr auto name = "json_extract";

-    static rapidjson::Value parse_json(const ColumnString* json_col, const ColumnString* path_col,
-                                       rapidjson::Document::AllocatorType& allocator,
-                                       const int row) {
+    static std::pair<bool, rapidjson::Value> parse_json(
+            const ColumnString* json_col, const ColumnString* path_col,
+            rapidjson::Document::AllocatorType& allocator, const int row) {
        rapidjson::Value value;
        rapidjson::Document document;

@ -860,11 +828,14 @@ struct FunctionJsonExtractImpl {
        const auto path = path_col->get_data_at(row);
        std::string_view path_string(path.data, path.size);

-        auto root = get_json_object<JSON_FUN_STRING>(json_string, path_string, &document);
+        auto* root = get_json_object<JSON_FUN_STRING>(json_string, path_string, &document);
+        bool found = false;
        if (root != nullptr) {
+            found = true;
            value.CopyFrom(*root, allocator);
        }
-        return value;
+
+        return {found, std::move(value)};
    }

    static void execute(const std::vector<const ColumnString*>& data_columns,
@ -874,30 +845,41 @@ struct FunctionJsonExtractImpl {
        rapidjson::StringBuffer buf;
        rapidjson::Writer<rapidjson::StringBuffer> writer(buf);

-        const auto json_col = data_columns[0];
+        const auto* json_col = data_columns[0];
        for (size_t row = 0; row < input_rows_count; row++) {
            rapidjson::Value value;
            if (data_columns.size() == 2) {
-                value = parse_json(json_col, data_columns[1], allocator, row);
+                auto result = parse_json(json_col, data_columns[1], allocator, row);
+                if (result.first) {
+                    value = std::move(result.second);
+                } else {
+                    null_map[row] = 1;
+                    result_column.insert_default();
+                    continue;
+                }
            } else {
+                bool found_any = false;
                value.SetArray();
                value.Reserve(data_columns.size() - 1, allocator);
                for (size_t col = 1; col < data_columns.size(); ++col) {
-                    value.PushBack(parse_json(json_col, data_columns[col], allocator, row),
-                                   allocator);
+                    auto result = parse_json(json_col, data_columns[col], allocator, row);
+                    if (result.first) {
+                        found_any = true;
+                        value.PushBack(std::move(result.second), allocator);
+                    }
+                }
+                if (!found_any) {
+                    null_map[row] = 1;
+                    result_column.insert_default();
+                    continue;
                }
            }

-            if (value.IsNull()) {
-                null_map[row] = 1;
-                result_column.insert_default();
-            } else {
-                // write value as string
-                buf.Clear();
-                writer.Reset(buf);
-                value.Accept(writer);
-                result_column.insert_data(buf.GetString(), buf.GetSize());
-            }
+            // write value as string
+            buf.Clear();
+            writer.Reset(buf);
+            value.Accept(writer);
+            result_column.insert_data(buf.GetString(), buf.GetSize());
        }
    }
 };
--- a/be/test/util/jsonb_parser_simd_test.cpp
+++ b/be/test/util/jsonb_parser_simd_test.cpp
@ -232,7 +232,7 @@ TEST_F(JsonbParserTest, ParseJsonWithLongInt) {

 TEST_F(JsonbParserTest, ParseInvalidJsonFormat) {
    std::string_view invalid_json = R"({"key": "value")";
-    EXPECT_EQ(parse_json_and_check(invalid_json, invalid_json), JsonbErrType::E_INVALID_KEY_STRING);
+    EXPECT_NE(parse_json_and_check(invalid_json, invalid_json), JsonbErrType::E_NONE);
 }

 TEST_F(JsonbParserTest, ParseJsonWithInvalidKeyType) {
--- a/be/test/vec/function/function_json_test.cpp
+++ b/be/test/vec/function/function_json_test.cpp
@ -72,15 +72,13 @@ TEST(FunctionJsonTEST, GetJsonStringTest) {
    std::string func_name = "get_json_string";
    InputTypeSet input_types = {TypeIndex::String, TypeIndex::String};
    DataSet data_set = {
-            {{VARCHAR("{\"k1\":\"v1\", \"k2\":\"v2\"}"), VARCHAR("$.k1")}, VARCHAR("v1")},
-            {{VARCHAR("{\"k1\":\"v1\", \"my.key\":[\"e1\", \"e2\", \"e3\"]}"),
-              VARCHAR("$.\"my.key\"[1]")},
+            {{VARCHAR(R"({"k1":"v1", "k2":"v2"})"), VARCHAR("$.k1")}, VARCHAR("v1")},
+            {{VARCHAR(R"({"k1":"v1", "my.key":["e1", "e2", "e3"]})"), VARCHAR("$.\"my.key\"[1]")},
             VARCHAR("e2")},
-            {{VARCHAR("{\"k1.key\":{\"k2\":[\"v1\", \"v2\"]}}"), VARCHAR("$.\"k1.key\".k2[0]")},
+            {{VARCHAR(R"({"k1.key":{"k2":["v1", "v2"]}})"), VARCHAR("$.\"k1.key\".k2[0]")},
             VARCHAR("v1")},
-            {{VARCHAR("[{\"k1\":\"v1\"}, {\"k2\":\"v2\"}, {\"k1\":\"v3\"}, {\"k1\":\"v4\"}]"),
-              VARCHAR("$.k1")},
-             VARCHAR("[\"v1\",\"v3\",\"v4\"]")}};
+            {{VARCHAR(R"([{"k1":"v1"}, {"k2":"v2"}, {"k1":"v3"}, {"k1":"v4"}])"), VARCHAR("$.k1")},
+             Null()}};

    static_cast<void>(check_function<DataTypeString, true>(func_name, input_types, data_set));
 }
--- a/regression-test/data/json_p0/test_json_load_and_function.out
+++ b/regression-test/data/json_p0/test_json_load_and_function.out
@ -6189,7 +6189,7 @@ false
 14	[123,456]	\N
 15	["abc","def"]	\N
 16	[null,true,false,100,6.18,"abc"]	\N
-17	[{"k1":"v41","k2":400},1,"a",3.14]	["v41"]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	"v31"
 26	\N	\N
 27	{"k1":"v1","k2":200}	"v1"
@ -6203,61 +6203,61 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,456]
-15	["abc","def"]	[null,"def"]
-16	[null,true,false,100,6.18,"abc"]	[null,true]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],1]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	[456]
+15	["abc","def"]	["def"]
+16	[null,true,false,100,6.18,"abc"]	[true]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	[1]
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,null]
-15	["abc","def"]	[null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],null]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 1	\N	\N
@ -6290,32 +6290,32 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null,null]
-3	true	[null,null,null]
-4	false	[null,null,null]
-5	100	[null,null,null]
-6	10000	[null,null,null]
-7	1000000000	[null,null,null]
-8	1152921504606846976	[null,null,null]
-9	6.18	[null,null,null]
-10	"abcd"	[null,null,null]
-11	{}	[null,null,null]
-12	{"k1":"v31","k2":300}	[null,null,null]
-13	[]	[null,null,null]
-14	[123,456]	[null,null,null]
-15	["abc","def"]	[null,null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[null,null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	\N
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	["v41",400,"a"]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[null,null,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null,null]
-29	12524337771678448270	[null,null,null]
-30	-9223372036854775808	[null,null,null]
-31	18446744073709551615	[null,null,null]
-32	{"":"v1"}	[null,null,null]
-33	{"":1,"":"v1"}	[null,null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null,null]
+27	{"k1":"v1","k2":200}	\N
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 2	null
@ -11664,7 +11664,7 @@ false
 14	[123,456]	\N
 15	["abc","def"]	\N
 16	[null,true,false,100,6.18,"abc"]	\N
-17	[{"k1":"v41","k2":400},1,"a",3.14]	["v41"]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	"v31"
 27	{"k1":"v1","k2":200}	"v1"
 28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
@ -11676,58 +11676,58 @@ false
 34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,456]
-15	["abc","def"]	[null,"def"]
-16	[null,true,false,100,6.18,"abc"]	[null,true]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],1]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	[456]
+15	["abc","def"]	["def"]
+16	[null,true,false,100,6.18,"abc"]	[true]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	[1]
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,null]
-15	["abc","def"]	[null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],null]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 2	null	\N
@ -11757,31 +11757,31 @@ false
 34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
-2	null	[null,null,null]
-3	true	[null,null,null]
-4	false	[null,null,null]
-5	100	[null,null,null]
-6	10000	[null,null,null]
-7	1000000000	[null,null,null]
-8	1152921504606846976	[null,null,null]
-9	6.18	[null,null,null]
-10	"abcd"	[null,null,null]
-11	{}	[null,null,null]
-12	{"k1":"v31","k2":300}	[null,null,null]
-13	[]	[null,null,null]
-14	[123,456]	[null,null,null]
-15	["abc","def"]	[null,null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[null,null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	\N
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	["v41",400,"a"]
-27	{"k1":"v1","k2":200}	[null,null,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null,null]
-29	12524337771678448270	[null,null,null]
-30	-9223372036854775808	[null,null,null]
-31	18446744073709551615	[null,null,null]
-32	{"":"v1"}	[null,null,null]
-33	{"":1,"":"v1"}	[null,null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null,null]
+27	{"k1":"v1","k2":200}	\N
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !sql_json_parse --
 {"":"v1"}
--- a/regression-test/data/jsonb_p0/test_jsonb_load_and_function.out
+++ b/regression-test/data/jsonb_p0/test_jsonb_load_and_function.out
@ -8187,7 +8187,7 @@ false
 14	[123,456]	\N
 15	["abc","def"]	\N
 16	[null,true,false,100,6.18,"abc"]	\N
-17	[{"k1":"v41","k2":400},1,"a",3.14]	["v41"]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	"v31"
 26	\N	\N
 27	{"k1":"v1","k2":200}	"v1"
@ -8201,61 +8201,61 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,456]
-15	["abc","def"]	[null,"def"]
-16	[null,true,false,100,6.18,"abc"]	[null,true]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],1]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	[456]
+15	["abc","def"]	["def"]
+16	[null,true,false,100,6.18,"abc"]	[true]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	[1]
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,null]
-15	["abc","def"]	[null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],null]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 1	\N	\N
@ -8288,32 +8288,32 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null,null]
-3	true	[null,null,null]
-4	false	[null,null,null]
-5	100	[null,null,null]
-6	10000	[null,null,null]
-7	1000000000	[null,null,null]
-8	1152921504606846976	[null,null,null]
-9	6.18	[null,null,null]
-10	"abcd"	[null,null,null]
-11	{}	[null,null,null]
-12	{"k1":"v31","k2":300}	[null,null,null]
-13	[]	[null,null,null]
-14	[123,456]	[null,null,null]
-15	["abc","def"]	[null,null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[null,null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	\N
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	["v41",400,"a"]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[null,null,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null,null]
-29	12524337771678448270	[null,null,null]
-30	-9223372036854775808	[null,null,null]
-31	18446744073709551615	[null,null,null]
-32	{"":"v1"}	[null,null,null]
-33	{"":1,"":"v1"}	[null,null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null,null]
+27	{"k1":"v1","k2":200}	\N
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !sql_json_length --
 1
--- a/regression-test/data/nereids_function_p0/scalar_function/J.out
+++ b/regression-test/data/nereids_function_p0/scalar_function/J.out
@ -7348,7 +7348,7 @@ false
 14	[123,456]	\N
 15	["abc","def"]	\N
 16	[null,true,false,100,6.18,"abc"]	\N
-17	[{"k1":"v41","k2":400},1,"a",3.14]	["v41"]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	"v31"
 26	\N	\N
 27	{"k1":"v1","k2":200}	"v1"
@ -7359,55 +7359,55 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,456]
-15	["abc","def"]	[null,"def"]
-16	[null,true,false,100,6.18,"abc"]	[null,true]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],1]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	[456]
+15	["abc","def"]	["def"]
+16	[null,true,false,100,6.18,"abc"]	[true]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	[1]
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,null]
-15	["abc","def"]	[null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],null]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N

 -- !select --
 1	\N	\N
@ -7437,29 +7437,29 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null,null]
-3	true	[null,null,null]
-4	false	[null,null,null]
-5	100	[null,null,null]
-6	10000	[null,null,null]
-7	1000000000	[null,null,null]
-8	1152921504606846976	[null,null,null]
-9	6.18	[null,null,null]
-10	"abcd"	[null,null,null]
-11	{}	[null,null,null]
-12	{"k1":"v31","k2":300}	[null,null,null]
-13	[]	[null,null,null]
-14	[123,456]	[null,null,null]
-15	["abc","def"]	[null,null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[null,null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	\N
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	["v41",400,"a"]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[null,null,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null,null]
-29	12524337771678448270	[null,null,null]
-30	-9223372036854775808	[null,null,null]
-31	18446744073709551615	[null,null,null]
+27	{"k1":"v1","k2":200}	\N
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N

 -- !json_extract_string --
 "v31"
--- a/regression-test/data/nereids_p0/json_p0/test_json_load_and_function.out
+++ b/regression-test/data/nereids_p0/json_p0/test_json_load_and_function.out
@ -6186,7 +6186,7 @@ false
 14	[123,456]	\N
 15	["abc","def"]	\N
 16	[null,true,false,100,6.18,"abc"]	\N
-17	[{"k1":"v41","k2":400},1,"a",3.14]	["v41"]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	"v31"
 26	\N	\N
 27	{"k1":"v1","k2":200}	"v1"
@ -6200,61 +6200,61 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,456]
-15	["abc","def"]	[null,"def"]
-16	[null,true,false,100,6.18,"abc"]	[null,true]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],1]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	[456]
+15	["abc","def"]	["def"]
+16	[null,true,false,100,6.18,"abc"]	[true]
+17	[{"k1":"v41","k2":400},1,"a",3.14]	[1]
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 1	\N	\N
-2	null	[null,null]
-3	true	[null,null]
-4	false	[null,null]
-5	100	[null,null]
-6	10000	[null,null]
-7	1000000000	[null,null]
-8	1152921504606846976	[null,null]
-9	6.18	[null,null]
-10	"abcd"	[null,null]
-11	{}	[null,null]
-12	{"k1":"v31","k2":300}	[300,null]
-13	[]	[null,null]
-14	[123,456]	[null,null]
-15	["abc","def"]	[null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[[400],null]
-18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	[300]
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
+18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	[300]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[200,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null]
-29	12524337771678448270	[null,null]
-30	-9223372036854775808	[null,null]
-31	18446744073709551615	[null,null]
-32	{"":"v1"}	[null,null]
-33	{"":1,"":"v1"}	[null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null]
+27	{"k1":"v1","k2":200}	[200]
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !select --
 1	\N	\N
@ -6287,32 +6287,32 @@ false

 -- !select --
 1	\N	\N
-2	null	[null,null,null]
-3	true	[null,null,null]
-4	false	[null,null,null]
-5	100	[null,null,null]
-6	10000	[null,null,null]
-7	1000000000	[null,null,null]
-8	1152921504606846976	[null,null,null]
-9	6.18	[null,null,null]
-10	"abcd"	[null,null,null]
-11	{}	[null,null,null]
-12	{"k1":"v31","k2":300}	[null,null,null]
-13	[]	[null,null,null]
-14	[123,456]	[null,null,null]
-15	["abc","def"]	[null,null,null]
-16	[null,true,false,100,6.18,"abc"]	[null,null,null]
-17	[{"k1":"v41","k2":400},1,"a",3.14]	[null,null,null]
+2	null	\N
+3	true	\N
+4	false	\N
+5	100	\N
+6	10000	\N
+7	1000000000	\N
+8	1152921504606846976	\N
+9	6.18	\N
+10	"abcd"	\N
+11	{}	\N
+12	{"k1":"v31","k2":300}	\N
+13	[]	\N
+14	[123,456]	\N
+15	["abc","def"]	\N
+16	[null,true,false,100,6.18,"abc"]	\N
+17	[{"k1":"v41","k2":400},1,"a",3.14]	\N
 18	{"k1":"v31","k2":300,"a1":[{"k1":"v41","k2":400},1,"a",3.14]}	["v41",400,"a"]
 26	\N	\N
-27	{"k1":"v1","k2":200}	[null,null,null]
-28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	[null,null,null]
-29	12524337771678448270	[null,null,null]
-30	-9223372036854775808	[null,null,null]
-31	18446744073709551615	[null,null,null]
-32	{"":"v1"}	[null,null,null]
-33	{"":1,"":"v1"}	[null,null,null]
-34	{"":1,"ab":"v1","":"v1","":2}	[null,null,null]
+27	{"k1":"v1","k2":200}	\N
+28	{"a.b.c":{"k1.a1":"v31","k2":300},"a":"niu"}	\N
+29	12524337771678448270	\N
+30	-9223372036854775808	\N
+31	18446744073709551615	\N
+32	{"":"v1"}	\N
+33	{"":1,"":"v1"}	\N
+34	{"":1,"ab":"v1","":"v1","":2}	\N

 -- !sql_json_parse --
 {"":"v1"}
--- a/regression-test/data/query_p0/sql_functions/json_functions/test_json_extract.out
+++ b/regression-test/data/query_p0/sql_functions/json_functions/test_json_extract.out
@ -20,3 +20,5 @@ v31
 -- !sql_string7 --
 \N

+-- !fix_array_path --
+[123]	[123]
--- a/regression-test/suites/query_p0/sql_functions/json_functions/test_json_extract.groovy
+++ b/regression-test/suites/query_p0/sql_functions/json_functions/test_json_extract.groovy
@ -28,4 +28,10 @@ suite("test_json_extract") {
        sql """ SELECT JSON_EXTRACT_STRING('{"id": 123, "name": "doris"}', '\$.'); """
        exception "Invalid Json Path for value: \$."
    }
+
+    qt_fix_array_path """
+        select 
+            JSON_EXTRACT('[{"key": [123]}]', '\$[0].key') v1
+            , JSON_EXTRACT('[{"key": [123]}]', '\$.[0].key') v2;
+    """
 }