How to validate that a number is int in C?

Question

how can i validate that the number which i scan is in integer in C? i wrote a code that works in every case other than when you enter a float/double like 5.00/6.00 or 13.00 . and almost every other answer i found ignores this situation. my solution so far was:

double tempValue;
printf("enter an integer:");
scanf("%lf",&tempValue);
int value= tempValue;
if(value==tempValue)
   printf("this is an int")
else printf("this isn't an integer");

thanks for reading. better focus on helping instead of down voting. (:

Answer 1

scan it into a double, then cast to int and check if there was a truncation:

double x = <your number>;
int y = (int) x;

if (x-y)
{
    // your number is NOT an integer
}
else
{
   // your number is an integer
}

Answer 2

In order to discriminate 5.00 and 5, you need to process the string.
Converting to a double with scanf loose this information.

For instance :

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

int main(void) {
        char  buffer[256];
        printf("enter an integer:");
        scanf("%255s",buffer);

        char  valueStr[256];
        sprintf(valueStr,"%d", atoi(buffer));
        if(strcmp(valueStr,buffer)==0)
                printf("'%s' is an integer\n",buffer);
        else
                printf("'%s' isn't an integer\n", buffer);

        return 0;
}

Live Test

This read the input as a string, convert it to an int, convert the int to a string and compare the input string with it.

Answer 3

#include <stdio.h>
#include <stdlib.h>

int main(void)
{
    int i, n, ch, done;

    n = scanf("%d", &i);
    if (n == 1) {
        ch = getchar();
        done = ch != '.';
        ungetc(ch, stdin);
    } else {
        done = 0;
    }
    if (done) {
        puts("this is an int");
    } else {
        puts("this is not an int");
    }
    return 0;
}

Answer 4

scanf() is made for formatted inputs (this is why it has an f in its name ^^). Here, your inputs are not formatted so you have to do something else.

The best way is to get the user input with fgets() and then to check the content. As mention is a previous comment, strtol() is the dedicated function for this task: you try to convert the string to a number and see if the conversion succeeds... or not.

Here is an example:

#include <assert.h>
#include <stdio.h>
#include <stdlib.h>

#define SIZE    1024

void convert() {

    // Get user input
    char buffer[SIZE] = {0};
    printf("Enter number: ");
    char* result = fgets(buffer, SIZE, stdin);
    assert(result != NULL);

    printf("User input = %s\n", buffer);    

    // Convert
    char* end = NULL;
    long int value = strtol(buffer, &end, 10);

    if (buffer[0] == '\n') {
        // The user didn't input anything
        // and just hit 'enter' kit
        puts("Nothing to convert");
    }    
    else if(*end == '\n') {
        // Assuming that 'buffer' was big enough to get all characters,
        // then the last character in 'buffer' is '\n'.
        // Because we want all characters except '\n' to be converted
        // we have to check that 'strtol()' has reached this character.
        printf("Conversion with success: %ld\n", value);
    } else {
        printf("Conversion failed. First invalid character: %c\n", *end);
    }
}

int main() {
  while(1) convert();
}

If you really want an int and not a long int, then you have to trick a little. My first idea would be to do something like this:

// Fits in integer?
int valueInt = value;
// some bits will be lost if 'value' doesn't fit in an integer
// hence the value will be lost

if (valueInt == value) {
  printf("Great: %ld == %d\n", value, valueInt);
} else {
  printf("Sorry: %ld != %d\n", value, valueInt);
}

Answer 5

Considering that you will never be certain that a arbitrary number is an integer (unless you check the input and treat it as a string).

The best approximation would be to check that the value compare to supposed the integer value would be lower than the machine epsilon

bool doubleMaybeInteger(double input) {
    return ((input - floor(input)) < DBL_EPSILON)
}