/root/doris/be/src/exprs/json_functions.h
Line | Count | Source (jump to first uncovered line) |
1 | | // Licensed to the Apache Software Foundation (ASF) under one |
2 | | // or more contributor license agreements. See the NOTICE file |
3 | | // distributed with this work for additional information |
4 | | // regarding copyright ownership. The ASF licenses this file |
5 | | // to you under the Apache License, Version 2.0 (the |
6 | | // "License"); you may not use this file except in compliance |
7 | | // with the License. You may obtain a copy of the License at |
8 | | // |
9 | | // http://www.apache.org/licenses/LICENSE-2.0 |
10 | | // |
11 | | // Unless required by applicable law or agreed to in writing, |
12 | | // software distributed under the License is distributed on an |
13 | | // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
14 | | // KIND, either express or implied. See the License for the |
15 | | // specific language governing permissions and limitations |
16 | | // under the License. |
17 | | |
18 | | #pragma once |
19 | | |
20 | | #include <fmt/format.h> |
21 | | #include <rapidjson/document.h> |
22 | | #include <simdjson.h> // IWYU pragma: keep |
23 | | |
24 | | #include <sstream> |
25 | | #include <string> |
26 | | #include <utility> |
27 | | #include <vector> |
28 | | |
29 | | #include "common/status.h" |
30 | | |
31 | | namespace simdjson { |
32 | | namespace fallback { |
33 | | namespace ondemand { |
34 | | class object; |
35 | | class value; |
36 | | } // namespace ondemand |
37 | | } // namespace fallback |
38 | | } // namespace simdjson |
39 | | |
40 | | namespace doris { |
41 | | |
42 | | enum JsonFunctionType { |
43 | | JSON_FUN_INT = 0, |
44 | | JSON_FUN_DOUBLE, |
45 | | JSON_FUN_STRING, |
46 | | |
47 | | JSON_FUN_UNKNOWN //The last |
48 | | }; |
49 | | |
50 | | struct JsonPath { |
51 | | std::string key; // key of a json object |
52 | | int idx; // array index of a json array, -1 means not set, -2 means * |
53 | | bool is_valid; // true if the path is successfully parsed |
54 | | bool is_reverse_index = false; // true if the path is last index, like '$.a[LAST(1)]' |
55 | | |
56 | | JsonPath(std::string key_, int idx_, bool is_valid_, bool is_reverse_index_) |
57 | | : key(std::move(key_)), |
58 | | idx(idx_), |
59 | | is_valid(is_valid_), |
60 | 54 | is_reverse_index(is_reverse_index_) {} |
61 | | |
62 | | JsonPath(std::string key_, int idx_, bool is_valid_) |
63 | 36 | : JsonPath(std::move(key_), idx_, is_valid_, false) {} |
64 | | |
65 | 0 | std::string to_string() const { |
66 | 0 | std::stringstream ss; |
67 | 0 | if (!is_valid) { |
68 | 0 | return "INVALID"; |
69 | 0 | } |
70 | 0 | if (!key.empty()) { |
71 | 0 | ss << key; |
72 | 0 | } |
73 | 0 | if (idx == -2) { |
74 | 0 | ss << "[*]"; |
75 | 0 | } else if (is_reverse_index) { |
76 | 0 | if (idx > 0) { |
77 | 0 | ss << "[last-" << idx << "]"; |
78 | 0 | } else { |
79 | 0 | ss << "[last]"; |
80 | 0 | } |
81 | 0 | } else if (idx > -1) { |
82 | 0 | ss << "[" << idx << "]"; |
83 | 0 | } |
84 | 0 | return ss.str(); |
85 | 0 | } |
86 | | |
87 | 0 | std::string debug_string() const { |
88 | 0 | return fmt::format("key:{}, idx:{}, valid:{}, is_reverse_index:{}", key, idx, is_valid, |
89 | 0 | is_reverse_index); |
90 | 0 | } |
91 | | }; |
92 | | |
93 | | class JsonFunctions { |
94 | | public: |
95 | | /** |
96 | | * The `document` parameter must be has parsed. |
97 | | * return Value Is Array object |
98 | | * wrap_explicitly is set to true when the returned Array is wrapped actively. |
99 | | */ |
100 | | static rapidjson::Value* get_json_array_from_parsed_json( |
101 | | const std::vector<JsonPath>& parsed_paths, rapidjson::Value* document, |
102 | | rapidjson::Document::AllocatorType& mem_allocator, bool* wrap_explicitly); |
103 | | |
104 | | // this is only for test, it will parse the json path inside, |
105 | | // so that we can easily pass a json path as string. |
106 | | static rapidjson::Value* get_json_array_from_parsed_json( |
107 | | const std::string& jsonpath, rapidjson::Value* document, |
108 | | rapidjson::Document::AllocatorType& mem_allocator, bool* wrap_explicitly); |
109 | | |
110 | | static rapidjson::Value* get_json_object_from_parsed_json( |
111 | | const std::vector<JsonPath>& parsed_paths, rapidjson::Value* document, |
112 | | rapidjson::Document::AllocatorType& mem_allocator); |
113 | | |
114 | | static void parse_json_paths(const std::string& path_strings, |
115 | | std::vector<JsonPath>* parsed_paths); |
116 | | // extract_from_object extracts value from object according to the json path. |
117 | | // Now, we do not support complete functions of json path. |
118 | | // Eg. city[*].id is not supported in this function |
119 | | static Status extract_from_object(simdjson::ondemand::object& obj, |
120 | | const std::vector<JsonPath>& jsonpath, |
121 | | simdjson::ondemand::value* value) noexcept; |
122 | | // src: {"a" : "b" {"c" : 1}, "e" : 123} |
123 | | // dst: {"a" : "b" {"d" : 1}} |
124 | | // merged: {"a" : "b" : {"c" : 1, "d" : 1}, "e" : 123} |
125 | | static void merge_objects(rapidjson::Value& dst_object, rapidjson::Value& src_object, |
126 | | rapidjson::Document::AllocatorType& allocator); |
127 | | |
128 | | static std::string print_json_value(const rapidjson::Value& value); |
129 | | |
130 | | static bool is_root_path(const std::vector<JsonPath>& json_path); |
131 | | |
132 | | private: |
133 | | static rapidjson::Value* match_value(const std::vector<JsonPath>& parsed_paths, |
134 | | rapidjson::Value* document, |
135 | | rapidjson::Document::AllocatorType& mem_allocator, |
136 | | bool is_insert_null = false); |
137 | | static void get_parsed_paths(const std::vector<std::string>& path_exprs, |
138 | | std::vector<JsonPath>* parsed_paths); |
139 | | }; |
140 | | } // namespace doris |