I have been doing coding Katas recently on codingame.com and codewars.com and something that i have noticed is that c++ doesn't have very good built in support for string splitting.
This is not that surprising but i find it unfortunate that with the changes they are making to the c++std there isn't a native way to easily do this perhaps something as simple as:
std::vector<iterator> spaces = std::find_all(begin(s), end(s), ' ');Well, in case you wanted to know some ways of doing this in c++, i have made a list of ways you could do this, please do comment other ways you have thought of. all of which will output the following to the console:
std::stringstream
#include <string>
#include <sstream>
#include <iostream>
int main()
{
using namespace std;
string a{"this is a string of text to split"};
stringstream stream(a);
string split;
int i{1};
while (getline(stream, split, ' '))
{
cout << i << ") " << split << endl;
++i;
}
}std::regex
#include <string>
#include <regex>
#include <iostream>
int main()
{
using namespace std;
string a{"this is a string of text to split"};
regex whitespace(R"(\s+)");
smatch match;
string buffer{a};
int i{1};
while (regex_search(buffer, match, whitespace))
{
cout << i << ") " << match.prefix() << endl;
buffer = match.suffix();
++i;
}
cout << i << ") " << buffer << endl;
}std::string.find()
#include <string>
#include <iostream>
int main()
{
using namespace std;
string a{"this is a string of text to split"};
size_t begin{0};
auto pos{a.find(' ', begin)};
int i{1};
while (pos != string::npos)
{
cout << i << ") " << a.substr(begin, pos - begin) << endl;
begin = pos + 1;
++i;
pos = a.find(' ', begin);
}
cout << i << ") " << a.substr(begin) << endl;
}I think those are the main ones. stringstream seems to be the least verbose. but which is faster?
I did a basic benchmark, i ran all these methods 1 million times and timed them the results are as follows:
MinGW x64 windows (Debug binaries):
std::stringstream took 1671ms
std::regex took 15994.1ms
std::string.find() took 1189.95ms
MinGW x64 windows (Release binaries):
std::stringstream took 907.998ms
std::regex took 3199.08ms
std::string.find() took 366.002ms
MSVC 14 (2017) x64 (Release binaries)
std::stringstream took 1117.02ms
std::regex took 6823.93ms
std::string.find() took 536.398ms
MSVC 14 (2017) x64 (Debug binaries)
std::stringstream took 74282.3ms
std::regex took 660995ms
std::string.find() took 76841.8ms
MSVC in debug binaries was so horrifyingly slow that i did it 1.5 times instead of 3 like the others.
But as you can see string.find() is consistently faster than the other methods, but because of stringstream being much less verbose, I think I will default to using that for readability.
Kommentare