How to name regex group matches in C++ the way python does (?P(.*))

Question

I have a string in my program that contains certain values for parameters. I need to extract the values from the parameters using regex.

The regex looks like this:

std::smatch param;
std::string str = "--name=AName --age=AnAge --gender=AGender"

if (std::regex_match(str, param, std::regex(".*--name=(\\w+) .*--age=(\\d+) .*--gender=(\\w+) .*"))) 
{
    //if it finds the order of the regex will come here and the values for each will be stored in param[1-3] 
}

The problem is the order of the params can come in different orders, for example:

std::string str = "--gender=AGender --name=AName --age=AnAge"
std::string str = "--age=AnAge --gender=AGender --name=AName"
std::string str = "--name=AName --gender=AGender --age=AnAge "

Is there a way to express in a single regex expression to be able to capture values despite of the order instead of doing on regex per parameter I want to find? If so how can I access such value? In python is possible to add an <id> before the desired group to then later access it using same identifier. In my example code I do that using smatch type variable but the access to it depends on the order that the string has and I cannot rely on that.

Looks like you are parsing program arguments. This means that you are using wrong tool for this task. You should use some program argument parsing library not Regular Expressions. — Marek R, Sep 26 '22 at 15:18
Look here. https://stackoverflow.com/questions/24656131/regex-for-existence-of-some-words-whose-order-doesnt-matter — cwbusacker, Sep 26 '22 at 15:36

cwbusacker · Answer 1 · 2022-09-26T15:54:27.803

Use this regex: "^(?=.*--name=(\\w+))(?=.*--age=(\\d+))(?=.*--gender=(\\w+)).+"

The one problem you'll run into is the fact that params won't be able to determine which item belongs to which parameter.

The way I would solve this problem would be to use std::string::find.

For example:

std::string str = "--name=AName --age=AnAge --gender=AGender";
size_t namePos = str.find("--name=");
size_t agePos = str.find("--age="); 
size_t genderPos = str.find("--gender=");

std::string name = "";
std::string gender = "";
std::string age = "";

if(namePos != std::string::npos)
{
  // Add 7 to namePos since the size of "--name=" is 7.
  // Assuming that the delimiter of the name is whitespace so find the first
  // whitespace after --name=
  name = str.substr(namePos + 7, str.find_first_of(" \n\r", namePos + 7) - (namePos + 7));
}

if(agePos != std::string::npos)
{  
  // Add 6 to agePos since the size of "--age=" is 6.
  // Assuming that the delimiter of the age is whitepace so find the first 
  // whitespace after --age=
  age = str.substr(agePos + 6, str.find_first_of(" \n\r", agePos + 6) - (agePos + 6));
}

if(genderPos != std::string::npos)
{
  // Add 9 to genderPos since the size of "--gender=" is 9.
  // Assuming that the delimiter of the gender is whitespace so find the first
  // whitespace after --gender=
  gender = str.substr(genderPos + 9, str.find_first_of(" \n\r", genderPos + 9) - (genderPos + 9));

std::cout << name << " " <<  gender << " " << age << std::endl;
}

Output:

AName AGender AnAge

score 0 · Answer 2 · answered Sep 26 '22 at 17:20

There are better tools to parse commandlines, but if you really want to use regex, you will find that Boost::Regex makes this much easier than the std::regex.

In particular, it supports named groups (see e.g. Boost Regular Expression: Getting the Named Group) which is the feature you request in your question title.

You can combine that with BOOST_REGEX_MATCH_EXTRA to keep all matches for all named groups (by default, only the last match for each capture group is accessible after the search.)

Then you can just make a big disjunction ((?<group1>...)|(?<group2>...)|...) in your regex for all the groups you may encounter, and you will be able to get all values out regardless of their order.

How to name regex group matches in C++ the way python does (?P(.*))

2 Answers2