What is the modern equivalent (C++) style for the

问题:

What is the best option if I want to "upgrade" old C-code to newer C++ when reading a file with a semicolon delimiter:

/* reading in from file C-like: */
fscanf(tFile, "%d", &mypost.nr); /*delimiter ; */
fscanf(tFile, " ;%[^;];", mypost.aftername);/* delimiter ; */
fscanf(tFile, " %[^;]", mypost.forename);   /*delimiter ; */
fscanf(tFile, " ;%[^;];", mypost.dept);/*delimiter ; */
fscanf(tFile, " %[^;];", mypost.position);/* delimiter ; */
fscanf(tFile, "%d", &mypost.nr2);

//eqivalent best C++ method achieving the same thing?

回答1:

You could overload the right-shift operator on istream for your struct, so:

std::istream& operator>>(std::istream& is, mypost_struct& mps) {
    is >> mps.nr;
    is.ignore(1, ';');
    is.getline(mps.forename, 255, ';');
    is.getline(mps.aftername, 255, ';');
    is >> mps.dept;
    is.ignore(1, ';');
    is >> mps.position;
    is.ignore(1, ';');
    is >> mps.nr2;

    return is;
}

Subsequently, input is as simple as is >> mypost;, where is is the file that you have opened.

Edit: @UncleBens Thanks for pointing this out, I had forgotten to take spaces in account. I have updated the answer, assuming that forename and aftername are likely to contain spaces. And there was this rather embarrasing bit about the delimiters being double-quoted...

I just checked it using a struct definition as under:

struct mypost_struct {
    int nr;
    char forename[255], aftername[255];
    int dept, position, nr2;
};

.. and the result was as expected.

回答2:

As @susmits says, but you can also use the returned stream as a conditional, like:

if (is >> mps.nr && is.ignore(1, ";") && is >> mps.aftername && ...) {
   // all is well ...
} else {
   // bad input format
}

or even:

if (is >> mps.nr >> ignore(";") >> mps.aftername >> ...) {
    // all is well ...
} else {
    // bad input format
}

回答3:

What is the best option if I want to "upgrade" old C-code to newer C++...?

IMHO, the best way to do this would be to read the file line-by-line and use regular expressions for parsing.