How do You Split a String in C++ Using a String Delimiter?

Dean R.

—December 15, 2022

The ProblemJump To Solution

Imagine you’re building a web service that requires you to store each word from a URL query. Your input string will have spaces represented by their HTML encoding, %20, like so:

#include <iostream>
#include <vector>

using namespace std;

vector<string> getWords(string s) {
    //YOUR CODE HERE
}

int main() {
    string s = "This%20is%20a%20string.";
    for (string x: getWords(s)) {
        cout << x << endl;
    }
    return 0;
}

There is no standard split() function in C++ like there is in other languages such as Python and JavaScript, so how can you split a string in C++ using a string delimiter?

The Solution

There are several fairly simple methods one can implement to solve this problem, each with pros and cons.

The getline() function is a simple and more readable solution whenever you need to split a string by a character and not a string of length > 1.

The strtok() function uses a static buffer to enable a fairly simple (albeit slightly less intuitive) implementation. Unlike getline(), strtok() supports a delimiter of type string, which seems to make this function a viable solution. However, while a string can be supplied as the delimiter to strtok(), the function interprets it as meaning “split the input by every character in the delimiter string“.

Since the question specifies that the delimiter must itself be a string, the recommended way to approach this problem would be to use string.find() and string.substr().

Use string.find() and string.substr()

For this method, we need to include the string library:

#include <string>

The string.find() function has the following signature:

string.find(string delim)

It returns the index of the first character of the delimiter within the string from which find() is called.

The string.substr() function has the following signature:

string.substr(size_t begin, size_t length)

It returns the substring of string that begins at index begin and has length length.

These two functions can be used together, like so:

#include <string>

vector<string> getWords(string s){
    vector<string> res;
    int pos = 0;
    while(pos < s.size()){
        pos = s.find("%20");
        res.push_back(s.substr(0,pos));
        s.erase(0,pos+3); // 3 is the length of the delimiter, "%20"
    }
    return res;
}

This
is
a
string.

This method treats the delimiter as a proper string, so if the input string is:

string s = "This%20is%20a%20str2ing.";

Then the output will be:

This
is
a
str2ing.

Note here that, although we erase s from the input string at each iteration, we pass the string by value and thus make a copy that we can destroy without affecting the original string.

Build Your Own

Of course, if you have an objection to using prebuilt functions altogether, you could always build your own method for splitting a string. Here is one example of how you might implement it:

vector<string> getWords(string s) {
    vector<string> res;
    string delim = "%20";
    string token = "";
    for (int i = 0; i < s.size(); i++) {
        bool flag = true;
        for (int j = 0; j < delim.size(); j++) {
            if (s[i + j] != delim[j]) flag = false;
        }
        if (flag) {
            if (token.size() > 0) {
                res.push_back(token);
                token = "";
                i += delim.size() - 1;
            }
        } else {
            token += s[i];
        }
    }
    res.push_back(token);
    return res;
}

Output:

This
is
a
str2ing.