Previous: 11.2.3 String Operations: strcmp() and strcat()
Up: 11.2 Library String Functions
Next: 11.2.5 File I/O with Strings
Previous Page: 11.2.3 String Operations: strcmp() and strcat()
Next Page: 11.2.5 File I/O with Strings

11.2.4 String Conversion Functions

Besides the functions for manipulating strings discussed in the previous sections (and others not discussed, but presented in Appendix ), the standard library provides several functions for converting the character (ascii) information in a string to other data types such as integers or floats.

We will illustrate the use of one such function, atoi(), by modifying our function getint() that we wrote in Chapter . Recall, this function reads the next valid integer from the standard input character by character, skipping over any leading white space, converts the character sequence to an integer representation, and returns the integer value. The prototype for this function is:

int getint(void);

GETINT: Write a program that reads only a valid integer. If there is an error in entering an integer, it detects the error and allows the user to re-enter the data.

The program driver is quite simple; it calls the function getint() that returns a valid integer read from the standard input. The driver then prints the integer returned by the function. Here is the algorithm for getint():

initialize valid to False

    while not a valid string
         read a string s
         set valid to True if s represents a valid integer

         if valid
              return an integer represented by the string s
         else
              print an error message

if *s is '+' or '-'
         valid = digitstr(s + 1);
    else
         valid = digitstr(s);

If s is a valid digit string, the function returns an equivalent integer using the standard library function, atoi(). The call atoi(s) returns the integer represented by the string s. The function atoi() has the prototype (included in stdlib.h):

int atoi(STRING s);

To check if all characters in a string are digits, getint() uses the function digitstr(). The algorithm modules are combined and implemented in a program shown in Figure 11.7.

The driver gets an integer and prints it. The function getint() reads the next white space delimited string in the input using scanf(). If the first character is a unary operator, we check for digits string starting at the pointer s + 1; otherwise, we check starting at the pointer s. The flag, valid, stores the value returned by digitstr(). If valid is True, we use atoi() to return the integer represented by s; otherwise, we print a message prompting the user to re-enter the integer, flush any remaining characters on the input line, and read the new input. The flag valid is initialized to False, and the loop continues as long as valid remains False, i.e. as long as a valid integer is not entered.

The function digitstr() traverses the string until a NULL appears. If a non-digit is encountered anytime during traversal, it returns False; otherwise, at the end of traversal, it returns True. It uses the library function, isdigit() to check if a character is a digit.

A sample session is shown below:

• ***Valid Integer Input***
• Type an integer: 123e
• ***Error in input
• Re-enter your integer: =123
• ***Error in input
• Re-enter your integer: -+123
• ***Error in input
• Re-enter your integer: -123
• Integer is -123