Using strtok in c
Here is my take at a reasonably simple tokenize helper that
- stores results in a dynamically growing array
- null-terminating the array
- keeps the input string safe (strtok modifies the input string, which is undefined behaviour on a literal char[], at least I think in C99)
To make the code re-entrant, use the non-standard strtok_r
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
char** tokenize(const char* input)
{
char* str = strdup(input);
int count = 0;
int capacity = 10;
char** result = malloc(capacity*sizeof(*result));
char* tok=strtok(str," ");
while(1)
{
if (count >= capacity)
result = realloc(result, (capacity*=2)*sizeof(*result));
result[count++] = tok? strdup(tok) : tok;
if (!tok) break;
tok=strtok(NULL," ");
}
free(str);
return result;
}
int main ()
{
char** tokens = tokenize("test string.");
char** it;
for(it=tokens; it && *it; ++it)
{
printf("%s\n", *it);
free(*it);
}
free(tokens);
return 0;
}
Here is a strtok
-free reimplementation of that (uses strpbrk
instead):
char** tokenize(const char* str)
{
int count = 0;
int capacity = 10;
char** result = malloc(capacity*sizeof(*result));
const char* e=str;
if (e) do
{
const char* s=e;
e=strpbrk(s," ");
if (count >= capacity)
result = realloc(result, (capacity*=2)*sizeof(*result));
result[count++] = e? strndup(s, e-s) : strdup(s);
} while (e && *(++e));
if (count >= capacity)
result = realloc(result, (capacity+=1)*sizeof(*result));
result[count++] = 0;
return result;
}
Do you need to store them separately? Two pointers into a modified char array will yield two separate perfectly usable strings.
That is we transform this:
char str[] ="test string.";
Into this:
char str[] ="test\0string.";
^ ^
| |
char *s1 ----- |
char *s2 -----------
.
#include <stdio.h>
#include <string.h>
int main ()
{
char str[] ="test string.";
char *firstname = strtok(str, " ");
char *lastname = strtok(NULL, " ");
if (!lastname)
lastname = "";
printf("%s, %s\n", lastname, firstname);
return 0;
}
What about using strcpy
:
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#define MAX_NAMES 2
int main ()
{
char str[] ="test string.";
char *names[MAX_NAMES] = { 0 };
char *test;
int i = 0;
test = strtok (str," ");
while (test != NULL && i < MAX_NAMES)
{
names[i] = malloc(strlen(test)+1);
strcpy(names[i++], test);
test = strtok (NULL, " ");
}
for(i=0; i<MAX_NAMES; ++i)
{
if(names[i])
{
puts(names[i]);
free(names[i]);
names[i] = 0;
}
}
return 0;
}
It contains much clutter to maintain a complete program and clean its resources, but the main point is to use strcpy
to copy each token into its own string.