Why is my string extraction function using back referencing in regex not working as intended?

Question

Extraction Function

string extractStr(string str, string regExpStr) {
    regex regexp(regExpStr);
    smatch m;
    regex_search(str, m, regexp);
    string result = "";
    for (string x : m)
        result = result + x;
    return result;
}

The Main Code

#include <iostream>
#include <regex>

using namespace std;

string extractStr(string, string);

int main(void) {
    string test = "(1+1)*(n+n)";
    cout << extractStr(test, "n\\+n") << endl;
    cout << extractStr(test, "(\\d)\\+\\1") << endl;
    cout << extractStr(test, "([a-zA-Z])[+-/*]\\1") << endl;
    cout << extractStr(test, "([a-zA-Z])[+-/*]([a-zA-Z])") << endl;
    return 0;
}

The Output

String = (1+1)*(n+n)
n\+n = n+n
(\d)\+\1 = 1+11
([a-zA-Z])[+-/*]\1 = n+nn
([a-zA-Z])[+-/*]([a-zA-Z]) = n+nnn

If anyone could kindly point the error I've done or point me to a similar question in SO that I've missed while searching, it would be greatly appreciated.

`for (string x : m)` would not iterate over different matches, but about groups of current match so `m[0]` (whole match), and `m[1]`.. `m[n]` (group from `(`..`)`). — Jarod42, Dec 03 '21 at 09:27

score 1 · Accepted Answer · answered Dec 03 '21 at 05:31

Regexes in C++ don't work quite like "normal" regexes. Specialy when you are looking for multiple groups later. I also have some C++ tips in here (constness and references).

#include <cassert>
#include <iostream>
#include <sstream>
#include <regex>
#include <string>


// using namespace std; don't do this!
// https://stackoverflow.com/questions/1452721/why-is-using-namespace-std-considered-bad-practice

// pass strings by const reference
// 1. const, you promise not to change them in this function
// 2. by reference, you avoid making copies

std::string extractStr(const std::string& str, const std::string& regExpStr)
{
    std::regex regexp(regExpStr);
    std::smatch m;
    std::ostringstream os; // streams are more efficient for building up strings

    auto begin = str.cbegin();
    bool comma = false;

    // C++ matches regexes in parts so work you need to loop
    while (std::regex_search(begin, str.end(), m, regexp))
    {
        if (comma) os << ", ";
        os << m[0];
        comma = true;
        begin = m.suffix().first;
    }

    return os.str();
}

// small helper function to produce nicer output for your tests.
void test(const std::string& input, const std::string& regex, const std::string& expected)
{
    
    auto output = extractStr(input, regex);
    if (output == expected)
    {
        std::cout << "test succeeded : output = " << output << "\n";
    }
    else
    {
        std::cout << "test failed : output = " << output << ", expected : " << expected << "\n";
    }
}

int main(void)
{
    std::string input = "(1+1)*(n+n)";
    
    test(input, "n\\+n", "n+n");
    test(input, "(\\d)\\+\\1", "1+1");
    test(input, "([a-zA-Z])[+-/*]\\1", "n+n");
    
    return 0;
}

Why is my string extraction function using back referencing in regex not working as intended?

1 Answers1