Is there a function similar to awk for C++?

binarybob0001 · 04-15-2008, 03:31 PM

I have a program that need to scan a string using awk, but it's not script; it's a C++ application. Is there a function that can do that?

BrianK · 04-15-2008, 04:23 PM

awk is a language, not a function... so what function of awk are you trying to mimic? There are many ways to manipulate strings in C++. You can start by looking at the string class.

binarybob0001 · 04-15-2008, 04:36 PM

I'm trying to separate fields using the ":" as the separator. Such as,

Code:

awk -F":" ' { print $1 }' /etc/group

It just seems like a waste to have to recreate this functionality if it already exists.

ta0kira · 04-15-2008, 04:36 PM

You can use regular expressions with <regex.h>, but I don't know about the rest.
ta0kira

Tinkster · 04-15-2008, 04:36 PM

Quote:

Originally Posted by binarybob0001

I have a program that need to scan a string using awk, but it's not script; it's a C++ application. Is there a function that can do that?

If you're using Qt ... they have a very feature-rich
regex handler for strings.

ta0kira · 04-15-2008, 04:45 PM

Maybe you should just use popen with that exact command line?
ta0kira

Edit: I just saw what you're doing. Take a look at strsep.

osor · 04-15-2008, 07:21 PM

If you want to use C-style code, there you can tokenize a string with strtok(). If you want a C++-style analogue, you might have to use something from Boost (e.g., Boost::Tokenizer).

dmail · 04-15-2008, 08:25 PM

Here is a piece of code I posted to another site some time ago.

Code:

std::vector<std::string> split(std::string const& str, std::string const& delim)
{
   using std::string;
   string::size_type start(0);
   string::size_type end(string::npos);
   std::vector<string> sub;

   while( (end = str.find(delim,start)) != string::npos)
   {
      sub.push_back(str.substr(start, end - start));
      start = end + delim.size();
   }
   if(start < str.size() )sub.push_back(str.substr(start, str.size() - start));
   return sub;
}

edit:
Just for good measure I will post the same warning I did when it was originally posted: "This makes some assumptions such as the delimiter does not start the string." It should give you a good idea what needs to be done.

binarybob0001 · 04-19-2008, 07:54 PM

Thanks for all of your help. At least, I'm not the only one writing the same thing. Perhaps one day when I'm bored I'll write a library that does similar things as our favorite language awk.

lvv · 06-13-2011, 12:04 AM

There is SCC - awk like C++ snippet compiler. Your:

Code:

awk -F: '{print $1}' /etc/group

will be:

Code:

scc -nF: '$1' /etc/group

theNbomr · 06-13-2011, 01:49 PM

In Linux, there is libpcre; perl-compatible regular expressions. For details:

Code:

man pcre

I don't know whether it is standard fare on most distros. You may have to install it.

--- rod.

sundialsvcs · 06-14-2011, 06:41 AM

You should find that libpcre is there.

What you are really looking for is a C++ library class that uses it, so that "all of the arcane magick" is done for you.

... assuming that you really do have to write a C/C++ level program to do this, versus using the very handy and well-debugged C/C++ program (awk itself) that already exists... mmmmm??

bigearsbilly · 06-16-2011, 04:43 PM

a simple example of pcre:

Code:

[0]$ echo hello:old:chap | ./pcre :
hello
old
chap

Code:

#include <iostream>
#include <vector>
#include <string>
#include <pcre++.h>

using namespace std;
using namespace pcrepp;

int main(int argc, char ** argv) {


    string pat = (argc==2)? argv[1]:"\\s+";
    vector<string> v;
    string s;
    Pcre p(pat);

    while( getline(cin, s) ) {
        v = p.split(s);

        vector<string>::iterator i;
        for(i=v.begin();i!=v.end();i++) {
            cout << *i << endl;
        }
    }


    return 0;
}

NOTE: I've tested boost regex, regex.h and pcre and
they still get whipped by perl for speed.