Skip to content

rfl::Object

rfl::Object<...> behaves similarly to std::map<std::string, ...> and std::unordered_map<std::string, ...>.

It will be represented in JSON or other formats as an object:

auto bart = rfl::Object<std::string>();
bart["first_name"] = "Bart";
bart["last_name"] = "Simpson";
bart["town"] = "Springfield";

This is results in the following JSON string:

{"first_name":"Bart","last_name":"Simpson","town":"Springfield"}

It supports almost all of the normal features you would expect from such a container. For instance, you can iterate through it just like any other map:

for (const auto& [k, v]: bart) {
    std::cout << k << ": " << v << std::endl;
}

However, unlike these containers, the order of fields is preserved. It is also possible to have duplicate keys:

auto bart = rfl::Object<std::string>();
bart["first_name"] = "Bart";
bart["last_name"] = "Simpson";
bart["town"] = "Springfield";

// Note that you need the .insert(...) method,
// because operator[] would overwrite the
// existing field.
bart.insert("first_name", "Lisa");

This results in the following JSON string:

{"first_name":"Bart","last_name":"Simpson","town":"Springfield","first_name":"Lisa"}

There are three different ways of inserting fields:

  • operator[...] creates a new field, if a field of this name doesn't already exist, and then assigns it.
  • .insert(...) inserts a new field at the end, potentially creating duplicate field names. Much like std::map, it supports several types of inputs:
    • .insert(std::pair(key, value))
    • .insert(key, value)
    • .insert(begin, end), where begin and end are iterators of a container of key-value-pairs
  • .emplace(...) is an alias for .insert that exists primarily for reasons of compatability with standard containers.

There are three different ways of accessing fields:

  • operator[...] creates a new field, if a field of this name doesn't already exist.
  • .at(...) throws an exception, if field of this name doesn't exist.
  • .get(...) returns an rfl::Result wrapping the field, or an rfl::Error if the field doesn't exist.

Note that it is most efficient, if you access fields in the order that they were placed.