Extract decimal numbers from string in C results in segmentation fault - c

Question

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
int Extract(char input[], double output[])
{
int i, j, len;
i=0;
j=0;
len=0;
char s[50];
while(i<strlen(input)-1)
{
if(input[i]==' ') i++;
j=0;
s[0]='\0';
while(input[i]!=',')
{
if(input[i]==' ') i++;
s[j]=input[i];
i++;
j++;
}
s[j]='\0';
i++;
printf("%s - ", s);
output[len]=(double)atof(s);
printf("Element %d: %lf\n", len, output[len]);
len++;
}
printf("%d", len);
return len;
}
int main(){
char s[120]="0.1,0.35,0.05,0.1,0.15,0.05,0.2.";
double v[1000];
int len = Extract(s, v);
int i;
for(i=0; i<len; i++)
{
printf("%d: %lf\n", i, v[i]);
}
return 1;
}
I try to run this code but even if it compiles correctly I have stack errors, can anybody help me?
Note that the string is composed by some decimal numbers separated by commas and the string ends with a .
UPDATE: maybe there was some dirty in the folder but now I have an output:
Length: 32
0.1 - Element 0: 0.000000
0.35 - Element 1: 0.000000
0.05 - Element 2: 0.000000
0.1 - Element 3: 0.000000
0.15 - Element 4: 0.000000
0.05 - Element 5: 0.000000
Segmentation fault (core dumped)
Since I already made the thread can I still take advantage of your help for converting the string into double, since atof is converting into float and probably that's the reason why it prints all 0.0000?

Answer 1

I think I found the problem:
Your loop only check if the character is not ','. At the end of your input, you do not have a ',' character, instead you have a '.' which will lead to your loop going on forever, resulting in the segfault. You can fix it by changing the last character of your input into a ','.
You do not have the right format for your float output. If you need to only print two significant figures, then change your %lf into %.2lf.
By the way, you check for spaces in your input, but it doesn't look like your input has any spaces. Maybe take those checks out?
It all depends on how regulated your input is. If you can, process your input before you feed it into your function.
Let us know if that helps!

Answer 2

Here I am presenting 3 codes, where the first addresses the segfault issue(s); the second, a verbose commentary on the Extract function as presented; third, an example of the Extract function written in one of many possible improved forms.
The main take-aways should be to always guard against buffer (array) overruns in code and make friends with the debugger.
The peppering of printf's in the code suggests a debugger is not being used. Coding anything beyond the trivial (hello, world?) begs debugger knowledge. Understanding tools such as a debugger is as important as knowing the language.
I Hope this serves as a guide and maybe even some inspiration. Good luck with your coding adventures.
Here's the original code with minimum changes to fix the segfault (array overrun)
int Extract(char input[], double output[])
{
int i, j, len;
i = 0;
j = 0;
len = 0;
char s[50];
while (i<strlen(input) - 1)
{
if (input[i] == ' ') i++;
j = 0;
s[0] = '\0';
/* Primary bug fix; guard against input array overrun *and* check for separator */
while (input[i] && input[i] != ',')
{
if (input[i] == ' ') i++;
s[j] = input[i];
i++;
j++;
}
s[j] = '\0';
/* bug fix; guard against input array overrun when incrementing */
if (input[i]) {
i++;
}
printf("%s - ", s);
output[len] = (double)atof(s);
printf("Element %d: %lf\n", len, output[len]);
len++;
}
printf("%d", len);
return len;
}
Here's a critique of the original code.
int Extract(char input[], double output[])
{
/* Tedious variable declaration and initialization */
int i, j, len;
i = 0;
j = 0;
len = 0;
/* why not 70? or 420? */
char s[50];
/* This is an exceedingly expensive way to determine end of string. */
while (i<strlen(input) - 1)
{
/* Why test for space? There are no spaces in sample input.
This increment risks overrunning the input array (segfault)
*/
if (input[i] == ' ') i++;
j = 0;
s[0] = '\0';
/* no guard against input array overrun */
while (input[i] != ',')
{
/* Why test for space? There are no spaces in sample input.
This increment risks overrunning the input array (segfault)
*/
if (input[i] == ' ') i++;
s[j] = input[i];
i++;
j++;
}
s[j] = '\0';
/* Bug - no guard against input array overrun when incrementing i */
i++;
/* these print statements suggest someone is NOT using a debugger - major fail if so. */
printf("%s - ", s);
output[len] = (double)atof(s);
printf("Element %d: %lf\n", len, output[len]);
len++;
}
/* again, this is easily seen in the debugger. Use the debugger. */
printf("%d", len);
return len;
}
Lastly, an alternative Extract with some (cherry picked) conventions.
int Extract(double* output, const int output_max, const char* input, const char separator)
{
/* declare variables in the scope they're needed and ALWAYS give variables meaningful names */
int input_index = 0, output_count = 0;
/* Detect end of string and guard against overrunning output buffer */
while (input[input_index] && output_count < output_max)
{
const int BUFFER_MAX = 50;
/* let the compiler init buffer to 0 */
char buffer[BUFFER_MAX] = { 0 };
int buffer_index = 0;
/* accumulate values into buffer until separator or end of string encountered */
while (input[input_index] && input[input_index] != separator)
{
buffer[buffer_index++] = input[input_index++];
if (buffer_index == BUFFER_MAX) {
/* Overrun, cannot process input; exit with error code. */
return -1;
}
}
/* only convert buffer if it had accumulated values */
if (buffer_index) {
/* note atof will discard, say, a trailing period */
output[output_count++] = atof(buffer);
}
/* Guard against input_index increment causing an array overrun (possible segfault) */
if (input[input_index]) {
input_index++;
}
}
return output_count;
}
int main() {
const int OUTPUT_MAX = 1000;
const char separator = ',';
const char* input = "0.1 ,0.35,0.05,0.1,0.15,0.05,0.2.";
double output[OUTPUT_MAX];
const int num_elems = Extract(output, OUTPUT_MAX, input, separator);
/* print results to stdout */
if (num_elems == -1) {
fprintf(stdout, "\nElement too long to process error\n");
}
else {
fprintf(stdout, "\nTotal number of elements: %d\n\n", num_elems);
for (int i = 0; i < num_elems; i++) {
fprintf(stdout, "Element %d: %lf\n", i, output[i]);
}
}
return num_elems;
}
Good luck with your coding adventures.