How to use regex_replace() - c++

Question

I need to insert a backslash before certain special characters like(',",\,?) when they are present in a string.
I don't want to use boost or any other string functions. Preferably algorithms of c++.
#include <stdio.h>
#include <regex>
#include <bits/stdc++.h>
int main(){
std::string str;
std::cout <<"Enter the string : ";
std::getline(std::cin, str);
str=std::regex_replace(str, std::regex("\\"), "\\\\");
str=std::regex_replace(str, std::regex("\'"), "\\\'");
str=std::regex_replace(str, std::regex("\?"), "\\\?");
str=std::regex_replace(str, std::regex("\""), "\\\"");
std::cout<< str<<std::endl;
}
input: testing\"input"?
output:testing\\\"input\"\?
Error message:
terminate called after throwing an instance of 'std::regex_error'
what(): regex_error

Answer 1

This can be done in a very simple approach. You need to look up more documentation on regex. Without special flags it will use the std::ECMAScript syntax.
You can put all your search characters in a character class. So in [] brackets. Example:
R"(['"\?])"
Then, for the replace string, you need to read about std::regex_replace. In the "fmt"-string, you can use special characters for back referencing.
For example "$&" will give you a copy the complete match.
With that you program will be as simpe as
#include <iostream>
#include <regex>
int main()
{
std::string text{R"(one 'two' ?three? "four" \five\)"};
std::cout << std::regex_replace(text, std::regex(R"(['"\?])"), R"(\$&)") << "\n";
return 0;
}
The raw string R"(some_raw_string)" will help you with the somehow unreadable escape character orgies.

Answer 2

I need to insert a backslash before certain special characters like(',",\,?) when they are present in a string.
Ok sure, so the regex_replace function will definitely do that for you. The trap to watch out for in this case is literal escaping and the interpretation of special characters.
The first level here is the special characters in C++ for string literals. This mainly concerns the double-quote character to start and end string literals, and the backslash character used to escape special characters, or to encode non-alphanumeric characters.
The second level is the special characters as far as the regular expression engine is concerned, which has its own regular expression grammar. This is more complex than the string literals in the language.
So if you want to encode a special character for a regular string literal, you need to escape it once. If you want to encode a special character to pass it literally to the regex compiler, you need to escape it twice.
For example, if you type:
"abc\n"
then the backslash-n will be interpreted as a linefeed character, so gives the byte sequence (including null-termination):
{ 0x61, 0x62, 0x63, 0x0a, 0x00 }
So if you want the backslash to be interpreted literally, you have to escape it, thus:
"abc\\n"
which results in:
{ 0x61, 0x62, 0x63, 0x5c, 0x6e, 0x00 }
If you just want to print this string, you will get the expected results. But if you pass this string to the regex engine, it will see the fourth byte is the backslash and treat it specially, escaping or interpreting the following character. If this is not valid, it throws an exception - which is what you're seeing.
When dealing with regular expressions, I think it's easier to work with raw strings. This is a special way you can write a literal string so the compiler does no interpretation of the string contents. This means you can pass strings to the regex engine directly, and essentially skip to the second level.
This is a new feature of C++11, where you prefix the string with a capital-R and then enclose the string contests with parentheses and an optional delimiter string (which simply needs to be unique).
I have tweaked your program to work the way you describe, using raw strings:
//
// Build with minimum C++ language level of C++11, eg:
//
// c++ --std=c++11 -o ans ans.cpp
#include <iostream>
#include <regex>
int main (int argc, char* argv[])
{
std::string str;
std::cout << "Enter the string : ";
std::getline(std::cin, str);
str = std::regex_replace(str, std::regex(R"(\\)"), R"(\\)");
str = std::regex_replace(str, std::regex(R"(')"), R"(\')");
str = std::regex_replace(str, std::regex(R"(\?)"), R"(\?)");
str = std::regex_replace(str, std::regex(R"(\")"), R"(\")");
std::cout << str << std::endl;
return 0;
}
Here's a sample session, exercising all the symbols:
Enter the string : one 'two' ?three? "four" \five\
one \'two\' \?three\? \"four\" \\five\\