C++ Tokenize String

Question

I'm looking for a simple way to tokenize string input without using non default libraries such as Boost, etc.

For example, if the user enters forty_five, I would like to seperate forty and five using the _ as the delimiter.

Answer 1

To convert a string to a vector of tokens (thread safe):

std::vector<std::string> inline StringSplit(const std::string &source, const char *delimiter = " ", bool keepEmpty = false)
{
    std::vector<std::string> results;

    size_t prev = 0;
    size_t next = 0;

    while ((next = source.find_first_of(delimiter, prev)) != std::string::npos)
    {
        if (keepEmpty || (next - prev != 0))
        {
            results.push_back(source.substr(prev, next - prev));
        }
        prev = next + 1;
    }

    if (prev < source.size())
    {
        results.push_back(source.substr(prev));
    }

    return results;
}

Answer 2

You can use the strtok_r function, but read the man pages carefully so you understand how it maintains state.

Answer 3

Look at this tutorial, which is by far the best tutorial on tokenization that I have found so far. It covers the best practices in the implementation of different methods that include using getline() and find_first_of() in C++ std, and strtok() in C.