6

I have following string:

char str[] = "A/USING=B)";

I want to split to get separate A and B values with /USING= as a delimiter

How can I do it? I known strtok() but it just split by one character as delimiter.

Box Box Box Box
  • 5,094
  • 10
  • 49
  • 67
Ryo
  • 995
  • 2
  • 25
  • 41
  • 1
    You can use [`strchr`](http://en.cppreference.com/w/c/string/byte/strchr) to find the `'/'`, `'='` and `')'` characters, and get the substrings using e.g. [`strncpy`](http://en.cppreference.com/w/c/string/byte/strncpy) (or just plain [`strcpy`](http://en.cppreference.com/w/c/string/byte/strcpy) if you can modify the source string). – Some programmer dude Jan 22 '16 at 08:56
  • I just known my input string have exactly `/USING=` is unique , `A` and `B` just for example ,may be it include `\` or `=` – Ryo Jan 22 '16 at 09:00
  • 1
    Then how about [`strstr`](http://en.cppreference.com/w/c/string/byte/strstr) to find starting position of the sub-string – Some programmer dude Jan 22 '16 at 09:01
  • 1
    you can use strstr get position of /USING, and split based on this position – Kai Iskratsch Jan 22 '16 at 09:01
  • @Ryo, I got an answer to your question. This was also asked on another website. My answer uses `strtok` but still successfully splits it. I have even mentioned where the question was asked earlier. Check it out. – Box Box Box Box Jan 22 '16 at 09:23

5 Answers5

8

As others have pointed out, you can use strstr from <string.h> to find the delimiter in your string. Then either copy the substrings or modify the input string to split it.

Here's an implementation that returns the second part of a split string. If the string can't be split, it returns NULL and the original string is unchanged. If you need to split the string into more substrings, you can call the function on the tail repeatedly. The first part will be the input string, possibly shortened.

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

char *split(char *str, const char *delim)
{
    char *p = strstr(str, delim);

    if (p == NULL) return NULL;     // delimiter not found

    *p = '\0';                      // terminate string after head
    return p + strlen(delim);       // return tail substring
}

int main(void)
{
    char str[] = "A/USING=B";
    char *tail;

    tail = split(str, "/USING=");

    if (tail) {
        printf("head: '%s'\n", str);
        printf("tail: '%s'\n", tail);
    }

    return 0;
}
M Oehm
  • 28,726
  • 3
  • 31
  • 42
1

I known strtok() but it just split by one character as delimiter

Nopes, it's not.

As per the man page for strtok(), (emphasis mine)

char *strtok(char *str, const char *delim);

[...] The delim argument specifies a set of bytes that delimit the tokens in the parsed string. [...] A sequence of two or more contiguous delimiter bytes in the parsed string is considered to be a single delimiter. [...]

So, it need not be "one character" as you've mentioned. You can using a string, like in your case "/USING=" as the delimiter to get the job done.

Community
  • 1
  • 1
Sourav Ghosh
  • 133,132
  • 16
  • 183
  • 261
  • 1
    Answer in not valid, this will not split "U/USING=N" correctly – Sebastian Jan 22 '20 at 09:56
  • @Sebastian This was more about the usage of `strtok()` with the shown input. If the input varies, there are other methods, like `strstr()` which can be utilized. – Sourav Ghosh Jan 22 '20 at 10:28
  • 2
    You give the impression that using strtok is a good idea and the OP will assume that it will parse other strings that have /USING= in them. You should at least warn him that it will only work for his example string and not for any arbitrary string containing the substring /USING= – Jerry Jeremiah Apr 16 '20 at 09:27
  • @JerryJeremiah I missed to see your point. Can you elaborate? – Sourav Ghosh Apr 17 '20 at 04:17
1

Here is a little function to do this. It works exactly like strtok_r except that the delimiter is taken as a delimiting string, not a list of delimiting characters.

char *strtokstr_r(char *s, char *delim, char **save_ptr)
{
    char *end;
    if (s == NULL)
        s = *save_ptr;

    if (s == NULL || *s == '\0')
    {
        *save_ptr = s;
        return NULL;
    }

    // Skip leading delimiters.
    while (strstr(s,delim)==s) s+=strlen(delim);
    if (*s == '\0')
    {
        *save_ptr = s;
        return NULL;
    }

    // Find the end of the token.
    end = strstr (s, delim);
    if (end == NULL)
    {
        *save_ptr = s + strlen(s);
        return s;
    }

    // Terminate the token and make *SAVE_PTR point past it.
    memset(end, 0, strlen(delim));
    *save_ptr = end + strlen(delim);
    return s;
}
Sebastian
  • 178
  • 1
  • 7
0

Dude, this answer is only valid if the input is this one, if were "abUcd/USING=efgh" your algorithm doesn't work.

This answer is the only valid for me:

char *split(char *str, const char *delim)
{
    char *p = strstr(str, delim);

    if (p == NULL) return NULL;     // delimiter not found

    *p = '\0';                      // terminate string after head
    return p + strlen(delim);       // return tail substring
}

int main(void)
{
    char str[] = "A/USING=B";
    char *tail;

    tail = split(str, "/USING=");

    if (tail) {
        printf("head: '%s'\n", str);
        printf("tail: '%s'\n", tail);
    }

    return 0;
}
-1

See this. I got this when I searched for your question on google.

In your case it will be:

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

int main (int argc, char* argv [])
{
    char theString [16] = "abcd/USING=efgh";
    char theCopy [16];
    char *token;
    strcpy (theCopy, theString);
    token = strtok (theCopy, "/USING=");
    while (token)
    {
        printf ("%s\n", token);
        token = strtok (NULL, "/USING=");
    }

    return 0;
}

This uses /USING= as the delimiter.

The output of this was:

abcd                                                                                                                                                                                                                      
efgh 

If you want to check, you can compile and run it online over here.

Box Box Box Box
  • 5,094
  • 10
  • 49
  • 67