C++提取JSON文档所有值的键路径方案技术问询

Requirement #

Extract the set of key paths for each value in a JSON document, storing the final result in a std::vector<std::vector<std::string>> named allPaths. Each path is a vector of strings composed of JSON keys. Important: Do not include paths to array elements, only the path to the array itself.

Example #

Sample JSON Document

{ "shopper": { "Id": "4973860941232342", "Context": { "CollapseOrderItems": false, "IsTest": false } }, "SelfIdentifiersData": { "SelfIdentifierData": [ { "SelfIdentifierType": { "SelfIdentifierType": "111" } }, { "SelfIdentifierType": { "SelfIdentifierType": "2222" } } ] } }

Expected Result

std::vector<std::vector<std::string>> allPaths; // Result container
allPaths.push_back({"shopper"}); // Object path
allPaths.push_back({"shopper", "Id"}); // String value path
allPaths.push_back({"shopper", "Context"}); // Object path
allPaths.push_back({"shopper", "Context", "CollapseOrderItems"}); // Boolean value path
allPaths.push_back({"shopper", "Context", "IsTest"}); // Boolean value path
allPaths.push_back({"SelfIdentifiersData"}); // Object path
allPaths.push_back({"SelfIdentifiersData", "SelfIdentifierData"}); // Array path

Current Attempt #

I'm using RapidJson to parse the JSON and trying to extract paths via DFS recursive backtracking, but there's a flaw in the recursive logic that prevents the implementation from being complete. Any C++ JSON parsing library is acceptable for the solution.

Existing Code #

#include <string>
#include <vector>
#include "rapidjson-master/include/rapidjson/document.h"
using namespace rapidjson;

// RapidJson type name array for checking value types
static const char* kTypeNames[] = { "Null", "False", "True", "Object", "Array", "String", "Number" };

void dfs(std::vector<std::vector<std::string>> &result, std::vector<std::string> &currPath, const Value &currElement) {
    // Copy current path and add to result to avoid modification affecting stored paths
    std::vector<std::string> newQuery(currPath);
    result.push_back(newQuery);
    // If current element is not an object, stop traversing down
    if (currElement.GetType() != 3) {
        return;
    }
    // Iterate all key-value pairs of current object, recurse on child nodes and backtrack
    for (Value::ConstMemberIterator itr = currElement.MemberBegin(); itr != currElement.MemberEnd(); ++itr) {
        //printf("Type of member %s is %s\n", itr->name.GetString(), kTypeNames[itr->value.GetType()]);
        currPath.push_back(itr->name.GetString());
        dfs(result, currPath, itr->value);
        currPath.pop_back();
    }
}

std::vector<std::vector<std::string>> generateAllQueries(std::string &plainTextJson) {
    Document document;
    document.Parse(plainTextJson.c_str());
    std::vector<std::vector<std::string>> result;
    std::vector<std::string> currPath;
    dfs(result, currPath, document);
    return result;
}

Need Help #

I'm looking for a fix to the recursive logic in the DFS function, or any alternative feasible implementation in C++ that meets the requirement.

内容的提问来源于stack exchange，提问作者PlsWork