C++ regex finds only 1 sub match

Question

// Example program
#include <iostream>
#include <string>
#include <regex>
int main()
{
 std::string strr("1.0.0.0029.443");

    std::regex rgx("([0-9])");
    std::smatch match;

    if (std::regex_search(strr, match, rgx)) {
        for(int i=0;i<match.size();i++)
            std::cout << match[i] << std::endl;
    }
}

this program should write

but it writes

1
1

checked it here http://cpp.sh/ and on visual studio, both same results.

Why does it find only 2 matches and why are they same?

As I understand from answers here, regex search stops at first match and match variable holds the necessary (sub?)string value to continue(by repeating) for other matches. Also since it stops at first match, () charachters are used only for sub-matches within the result.

Dmitry Egorov · Accepted Answer · 2017-06-22T18:43:55.083

12

Being called once, regex_search returns only the first match in the match variable. The collection in match comprises the match itself and capture groups if there are any.

In order to get all matches call regex_search in a loop:

while(regex_search(strr, match, rgx))
{
    std::cout << match[0] << std::endl;
    strr = match.suffix();
}

Note that in your case the first capture group is the same as the whole match so there is no need in the group and you may define the regex simply as [0-9] (without parentheses.)

Demo: https://ideone.com/pQ6IsO

edited Jun 22 '17 at 18:43

answered Jun 22 '17 at 18:41

Dmitry Egorov

9,542
3
22
40

Why does it put only 1 search match into match variable? It stops searching at first match? – huseyin tugrul buyukisik Jun 22 '17 at 18:46
`\\d` is a slightly shorter alternative to `[0-9]` – AndyG Jun 22 '17 at 18:47
thank you for this important information. – huseyin tugrul buyukisik Jun 22 '17 at 18:47
2

@huseyintugrulbuyukisik: I think you're misunderstanding what the match variable is for. It is not for holding multiple matches. It is for holding matching sub-groups within the match (groups specified in parenthesis). – Benjamin Lindley Jun 22 '17 at 18:49
If I want to find three floating point variables, I need to loop for 3 times then, if I understand correctly. (ofcourse with proper regex code) – huseyin tugrul buyukisik Jun 22 '17 at 18:52

score 1 · Answer 2 · answered Jun 22 '17 at 18:49

Problems:

Using if only gives you one match. You need to use a while loop to find all the matches. You need to search past the previous match in the next iteration of the loop.
std::smatch::size() returns 1 + number of matches. See its documentation. std::smatch can contain sub-matches. To get the entire text, use match[0].

Here's an updated version of your program:

#include <iostream>
#include <string>
#include <regex>

int main()
{
   std::string strr("1.0.0.0029.443");

   std::regex rgx("([0-9])");
   std::smatch match;

   while (std::regex_search(strr, match, rgx)) {
      std::cout << match[0] << std::endl;
      strr = match.suffix();
   }
}

I guess there is no one-liner as in C# to get all parts with string.Split(stringArray,..) — huseyin tugrul buyukisik, Jun 22 '17 at 18:50
@huseyintugrulbuyukisik, I haven't programmed in C#. I can't respond to that. — R Sahu, Jun 22 '17 at 18:52

C++ regex finds only 1 sub match

2 Answers2