Ignoring numbers with letters around them - c

Question

#include <stdio.h>
#include <ctype.h>
int main(void)
{
const char *str = NULL;
size_t len = 0;
getline(&str,&len,stdin);
int i = 0;
unsigned int count = 0, tmp = 0;
printf("%s\n", str);
while (sscanf(&str[count], "%d %n", &i, &tmp) != EOF) {
count += tmp;
printf("number %d\n", i);
}
return 0;
}
What I'm trying to do here is ask the user for a line of integers (positive or negative) and I only care about integers that are valid that have a space between them, if there was a non numeric char entered then it should ignore what comes after it. Up top is what I have so far but I have some bugs with it, what I have right now is also ignoring white space which is what I want. Look at EXP for more clarification.
EXP:
input: 1 2 3 4a desired output-> 1 2 3 my ouput-> 1 2 3 4 4
input: 1 2 3 a4 desired output-> 1 2 3 my ouput-> 1 2 3 3
input: 1 2 3 a b c 4 5 6 7 desired output-> 1 2 3 my output-> 1 2 3 3 3 3 4 5 6 7
Valid input EXP:
input: 12 367 98 -3 -67 desired output-> 12 367 98 -3 -67

Answer 1

Whenever you need to work through a buffer picking out numeric values, you want to think strtoX() (where X can be l, ul, f, d, for strtol(), etc..) The reason being is two-fold. (1) these functions were designed to allow you to do that by providing parameters of a pointer to the string itself, and the endptr parameter that is updated to point to the next character after the last digit successfully converted; and (2) these function provide complete error detection for all aspects of the conversion.
By virtue of the endptr parameter being updated to one-past the last digit converted, you simply loop through your buffer, attempting the conversion, and upon successfully conversion, you just update the the initial pointer to endptr and repeat until you run out of characters. See man 3 strtol
Since endptr is updated to point one-past the last digit converted, you can simply check whether the character pointed to be endptr is either the nul-terminating character for the string or !isspace(*endptr) to determine if a non-digit was included in that group of numbers.
A simple implementation for your case would be:
#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
#include <errno.h>
#define MAXC 1024
int main (int argc, char **argv) {
char buf[MAXC]; /* buffer to hold each line */
/* use filename provided as 1st argument (stdin by default) */
FILE *fp = argc > 1 ? fopen (argv[1], "r") : stdin;
if (!fp) { /* validate file open for reading */
perror ("file open failed");
return 1;
}
while (fgets (buf, MAXC, fp)) { /* read each line into buf */
char *p = buf, *ep = p; /* pointers to use with strtol() */
for (;;) { /* loop continually */
errno = 0; /* reset errno */
long val = strtol (p, &ep, 0); /* attempt conversion to long */
if (p == ep) /* check if no characters converted */
break;
else if (errno) /* check for under/overflow */
break;
else if (*ep && !isspace(*ep)) /* check next char for non-space */
break;
else
printf (" %ld", val); /* good all-digit conversion */
p = ep; /* update pointer to endpointer */
}
putchar ('\n'); /* tidy up with newline */
}
if (fp != stdin) /* close file if not stdin */
fclose (fp);
}
Example Input File
$ cat dat/numwalpha.txt
1 2 3 4a
1 2 3 a4
1 2 3 a b c 4 5 6 7
Example Use/Output
$ ./bin/strtol_numonly dat/numwalpha.txt
1 2 3
1 2 3
1 2 3
Adding Error Reporting
On failure, you can simply output the characters beginning at endptr (ep above) to see why things failed. You can include a preprocessor directive to turn the error reporting on/off by using a DEBUG define if you like, e.g.
#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
#include <errno.h>
#define MAXC 1024
int main (int argc, char **argv) {
char buf[MAXC]; /* buffer to hold each line */
/* use filename provided as 1st argument (stdin by default) */
FILE *fp = argc > 1 ? fopen (argv[1], "r") : stdin;
if (!fp) { /* validate file open for reading */
perror ("file open failed");
return 1;
}
while (fgets (buf, MAXC, fp)) { /* read each line into buf */
char *p = buf, *ep = p; /* pointers to use with strtol() */
for (;;) { /* loop continually */
errno = 0; /* reset errno */
long val = strtol (p, &ep, 0); /* attempt conversion to long */
if (p == ep) { /* check if no characters converted */
#ifdef DEBUG
fprintf (stderr, "error: no digits converted in '%s'.\n", p);
#endif
break;
}
else if (errno) { /* check for under/overflow */
#ifdef DEBUG
fprintf (stderr, "error: over/under-flow in conversion or '%s'.\n", p);
#endif
break;
}
else if (*ep && !isspace(*ep)) { /* check next char for non-space */
#ifdef DEBUG
fprintf (stderr, "error: mixed alphanumeric '%s'.\n", ep);
#endif
break;
}
else
printf (" %ld", val); /* good all-digit conversion */
p = ep; /* update pointer to endpointer */
}
putchar ('\n'); /* tidy up with newline */
}
if (fp != stdin) /* close file if not stdin */
fclose (fp);
}
Then compiled with -DDEBUG to enable the define, you would have:
$ ./bin/strtol_numonly dat/numwalpha.txt
error: mixed alphanumeric 'a
'.
1 2 3
error: no digits converted in ' a4
'.
1 2 3
error: no digits converted in ' a b c 4 5 6 7
'.
1 2 3
Which shows the first case failed due to an 'a' at the end, the second due to an "a4" as the string to be converted, and the last because "a b c 4 5 6 7" was all that was left in the string, with 'a' being the next conversion.