将数组字符串转换为char

Question

i have taken all txt from file and place line by line into array of string. I am trying to split this string so that i can save word by word in separate array. Kindly just tell me how shell i convert the array of string into char.

for example

string line[15];  // line[0] has : was there before
                  // line[1] has : then after

char * pch;
char *c = line.c_str();   // string to char  (i am getting error here. Any body know?)
  pch = strtok (c," ");
  while (pch != NULL)
  {
    printf ("%s\n",pch);
    pch = strtok (NULL, " ");
  }

error: C2228: left of '.c_str' must have class/struct/union

Answer 1

string line[15]; is an array of strings. So when you have line.c_str(); line is a pointer to a string and not a string itself. A pointer doesn't have a .c_str() method and that's why the compiler is complaining. (Pointers don't have any methods and hence the compiler tells you that the left hand side of the expression must be a class/struct/union type). To fix this you want to index into the array to get a string. You can do this with something like: line[0].c_str();

Additionally you can't write to anything returned by c_str() as it returns a const pointer. So you'll need to copy the results from c_str first before you then operate on it if you are going to change it in place.

Also it might be worth mentioning that there's c++ ways of doing tokenizing, you might find some examples here Split a string in C++? . The last time I did this I was already using boost so I made use of the boost::tokenizer library.

Answer 2

There are simpler ways to accomplish this in C++. The strtok function is a C function and cannot be used with std::string directly since there is no way to get writable pointer access to the underlying characters in an std::string . You can use iostream s to get individual words separated by spaces from a file directly in C++.

Unfortunately the standard library lacks a simple, flexible, efficient method to split strings on arbitrary characters. It gets more complicated if you want to use iostream s to accomplish splitting on something other than whitespace. Using boost::split or the boost::tokenizer suggestion from shuttle87 is a good option if you need something more flexible (and it may well be more efficient too).

Here's a code example reading words from standard input, you can use pretty much the same code with an ifstream to read from a file or a stringstream to read from a string: http://ideone.com/fPpU4l

#include <algorithm>
#include <iostream>
#include <iterator>
#include <string>

using namespace std;

int main() {
    vector<string> words;
    copy(istream_iterator<string>{cin}, istream_iterator<string>{}, back_inserter(words));
    copy(begin(words), end(words), ostream_iterator<string>{cout, ", "});
    cout << endl;
}