Search code examples
cencodingcharascii

Please explain what this code is doing (someChar - 48)

I'm going through some practice problems, and I saw this code:

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

int main(void) {
   char* s = "357";
   int sum = 0;
   int i = 0;
   for (i = 0; i < strlen(s); i++) {
     sum += s[i] - 48;
   }
   printf("Sum is %d", sum);

   return 0;
}

Can someone explain what the code does, especially the subtraction with 48 part?

Solution

  • The code basically sums the digits of a number represented as a string. It makes two important assumptions to work properly:

    • The string contains only chars in the '0'..'9' range
    • The character encoding used is ASCII

    In ASCII, '0' == 48, '1' == 49, and so on. Thus, '0' - 48 == 0, '1' - 48 == 1, and so on. That is, subtracting by 48 translates the char values '0'..'9' to the int values 0..9.

    Thus, precisely because '0' == 48, the code will also work with:

    sum += s[i] - '0';

    The intention is perhaps slightly more clear in this version.

    You can of course do the "reverse" mapping by addition, e.g. 5 + '0' == '5'. Similarly, if you have a char containing a letter in 'A'..'Z' range, you can "subtract" 'A' from it to get the index of that letter in the 0..25 range.

    See also

    Related questions

    On alternative encodings

    As mentioned, the original - 48 code assumes that the character encoding used is ASCII. - '0' not only improves readability, but also waives the ASCII assumption, and will work with any encoding, as specified by the C language which stipulates that digit characters must be encoded sequentially in a contiguous block.

    On the other hand, no such stipulation is made about letters. Thus, in the rare situation where you're using EBCDIC encoding, for example, mapping 'A'..'Z' to 0..25 is no longer as simple as subtracting 'A', due to the fact that letters are NOT encoded sequentially in a contiguous block in EBCDIC.

    Some programming languages simplify matters by mandating one particular encoding is used to represent the source code (e.g. Java uses Unicode: JLS §3.1)

    See also

    Related questions