3 replies to this topic

[Kevn]

Hey, I was wondering if there is a better/more efficient way of going about extracting any number that follows 'i' in a file. So lets say the document has the following content : "Hello, this is a random document... try to extract i1 i34 i405 i2 from this document". So by the end of the extraction you would have the numbers 1, 34, 405, and 2.

The way I am currently doing it is:

string filename, line;
int test = 0;
ifstream infile(filename);
if (!infile){
	cout << "Could not open file" << endl;
	return 1;
}
while (getline(infile, line))
{
	for (string::iterator it = line.begin(); it < line.end(); it++)
	{
		 if (*it == 'i'){
			  it++;
			  test = 0;
			  while (*it > 47 && *it < 58){
					test = test * 10;
					test = test + (*it - '0');
					it++;
				}
				cout << test << " ";
				it--;
			}
	  }
}
infile.close();

[fastcall22]

ifstream fs( "test.txt" );
if ( !fs )
    return;

fs.seekg( 0, ios::end );
int sz = fs.tellg();
fs.seekg( 0, ios::beg );

char* file = new char[sz];
in.read( file, sz );

vector<int> data;
char* current = file;
char* previous = file;
char* end = file + sz;
int i;
while ( current < end ) {
    previous = current;
    i = strtol( current, &current, 10 );
    if ( previous == current )
        current ++;
    else
        data.push_back( i );
}

delete[] file;

</int>

EDIT:
Oops, didn't see that the integers needed to be prefixed with 'i'.

[SiCrane]

  while (infile) {

    std::istream::int_type c = infile.get();

    if (c == 'i') {

      std::istream::int_type next = infile.peek();

      if (isdigit(next)) {

        int i;

        infile >> i;

        std::cout << i << std::endl;

      }

    }

  }

[Kevn]

Awesome, thank you!