Correct way to use cin.fail()

Question

What is the correct way to use cin.fail();?

I am making a program where you need to input something. It is not very clear if you need to input a number or character. When a user inputs a character instead of a number the console goes crazy. How can I use cin.fail() to fix this?

Or is there a better way?

For a situation like this, you'd usually read a character, then check if it's a digit (and if so, convert to a number). — Jerry Coffin, Jul 29 '13 at 16:13
You could use a good C++ book. but have a look at: http://en.cppreference.com/w/cpp/io/basic_ios/fail — hetepeperfan, Jul 29 '13 at 16:15
@JerryCoffin Okay, I'll do that, but I would still like to learn to use `cin.fail()`. — , Jul 29 '13 at 16:16
@hetepeperfan I do have a good book. It just doesn't have anything about this. — , Jul 29 '13 at 16:20
@JerryCoffin I'd do the opposite. Getting things like skipping white space and reinserting the character in the stream right isn't obvious to a beginner. — James Kanze, Jul 29 '13 at 16:23

score 26 · Answer 1 · answered Jul 29 '13 at 16:22

std::cin.fail() is used to test whether the preceding input succeeded. It is, however, more idiomatic to just use the stream as if it were a boolean:

if ( std::cin ) {
    //  last input succeeded, i.e. !std::cin.fail()
}

if ( !std::cin ) {
    //  last input failed, i.e. std::cin.fail()
}

In contexts where the syntax of the input permit either a number of a character, the usual solution is to read it by lines (or in some other string form), and parse it; when you detect that there is a number, you can use an std::istringstream to convert it, or any number of other alternatives (strtol, or std::stoi if you have C++11).

It is, however, possible to extract the data directly from the stream:

bool isNumeric;
std::string stringValue;
double numericValue;
if ( std::cin >> numericValue ) {
    isNumeric = true;
} else {
    isNumeric = false;
    std::cin.clear();
    if ( !(std::cin >> stringValue) ) {
        //  Shouldn't get here.
    }
}

Okay, I never knew that was possible. Thank you! – Jul 29 '13 at 16:25 — , Jul 29 '13 at 16:25

score 12 · Accepted Answer · edited Oct 29 '18 at 09:07

12

cin.fail() returns true if the last cin command failed, and false otherwise.

An example:

int main() {
  int i, j = 0;

  while (1) {
    i++;
    cin >> j;
    if (cin.fail()) return 0;
    cout << "Integer " << i << ": " << j << endl;  
  }
}

Now suppose you have a text file - input.txt and it's contents are:

  30 40 50 60 70 -100 Fred 99 88 77 66

When you will run above short program on that, it will result like:

  Integer 1: 30
  Integer 2: 40
  Integer 3: 50
  Integer 4: 60
  Integer 5: 70
  Integer 6: -100

it will not continue after 6th value as it quits after reading the seventh word, because that is not an integer: cin.fail() returns true.

edited Oct 29 '18 at 09:07

pensono

336
6
17

answered Jul 29 '13 at 16:23

Shumail

3,103
4
28
35

2

@hetepeperfan numerically speaking 0 is the only false value. – Borgleader Jul 29 '13 at 16:33
Yeah but it will work in this situation - i edited answer and added `int main()` as well. – Shumail Jul 29 '13 at 16:34
Something's wrong here (in addition to the poor formatting). First, of course, `std::istream::fail()` never returns "non-zero" or "zero": it returns `true` or `false`. Second, on the given text file it outputs everything, **-100** is an integer, and `std::cin.fail()` will return `false` after reading it. – James Kanze Jul 29 '13 at 17:04
This fails when entering `1.2` or `1,2` – Peon Sep 26 '17 at 16:56
@JamesKanze "non-zero" and "zero" ARE true and false respectively in C++. – Layne Bernardo Sep 11 '21 at 01:34
@Peon That's because those aren't integers. – Layne Bernardo Sep 11 '21 at 01:35

Correct way to use cin.fail()

2 Answers2

Linked

Related