/^[^\s]+\s([^\s]+)\s/
In PHP,I can use regex to get the substr by $1,
how should I do it in C?
It's better if can do it without regex,though.
UPDATE
Put simply, how do I get werwerur out of swerwer werwerur y (the second)?
views:
119answers:
5
+2
A:
Use strtok() to find the spaces.
#include <string.h>
...
char *token;
char *line = "LINE TO BE SEPARATED";
char *search = " ";
/* Token will point to "LINE". */
token = strtok(line, search);
/* Token will point to "TO". */
token = strtok(NULL, search);
WhirlWind
2010-05-07 17:56:09
The string can be arbitary long,so `strtok()` may be inefficient.
httpinterpret
2010-05-07 17:58:50
@httpinterpret What do you *really* want? I am not a linear search magician.
WhirlWind
2010-05-07 17:59:59
I mean it's unnecessary to tokenize the whole string since I only care about the first two.
httpinterpret
2010-05-07 18:01:39
@httpinterpret then stop after the first two... strtok() only finds the next token.
WhirlWind
2010-05-07 18:02:15
Oh,it seems I misunderstood `strtok`,let me have a look again.
httpinterpret
2010-05-07 18:06:57
Strange,does `strtok` internally caches `line`(or remember history operations) so that `token = strtok(NULL, search); ` can point to `TO`?
httpinterpret
2010-05-07 18:09:04
@httpinterpret yes, each time you call strtok, it advances the value it returns to the next token. Thus, it is not thread safe.
WhirlWind
2010-05-07 18:10:27
There's a companion function called `strtok_r` that allows the caller to provide the position memory so that it can be thread safe if needed.
Tyler McHenry
2010-05-07 18:26:18
it's incorrect: char *line = "LINE TO BE SEPARATED"; <= should be const char*; correct: char line[] = "..."
oraz
2010-05-07 18:36:39
`strtok()` is up there as one of the worst functions in the C library. It modifies its input, so you may need to copy it. It keeps state in the function itself, so it's neither thread-safe nor re-entrant (if you try using it on a second string while tokenizing the first one, you've messed up). Try to find a safer function (which might be `strtok_r()`).
David Thornley
2010-05-07 18:37:08
A:
If I understand what you want correctly, sscanf should be able to do the job:
char result[256];
sscanf("swerwer werwerur y", "%*s %255s", result);
Jerry Coffin
2010-05-07 18:00:46
Since the matched length is unknown,so this kind of fixed length `result` may fail under certain cases.
httpinterpret
2010-05-07 18:03:38
A:
You could look for spaces in a char-string by using this:
char* strText = "word1 word2 word3";
int iTextLen = strlen(strText);
for(int i = 0; i < iTextLen; i++)
{
if(strText[i] == ' ') //is it a space-character
{
//save position or whatever
}
}
If you're looking for the second word, you just have to save the positions of the first space and of the second space and get the character string at the indices between the two.
Simon
2010-05-07 18:01:08
+1
A:
I recommend to use strchr() - it is very fast to find characters in strings
#include <string.h>
..
char str[] = "swerwer werwerur y";
char *p1 = NULL,*p2 = NULL;
p1 = strchr(str,' ');
p1++;
p2 = strchr(p1,' ');
if(p2) *p2 = 0;
printf("found: %s\n", p1);
if you have multiple delimiters, you can use strtok_r() or strpbrk() as in example below:
char str[] = "swerwer ., werwerur + y";
const char *dlms = " .,+";
char *p1 = NULL,*p2 = NULL;
p1 = strpbrk(str,dlms);
while(strchr(dlms,*p1)) p1++;
p2 = strpbrk(p1,dlms);
if(p2) *p2 = 0;
printf("found: %s\n", p1);
(should cleanup code: in case if strpbrk returns NULL)
oraz
2010-05-07 18:24:32
A:
If you are using a posix/unix system like Linux, you could use the regex API directly:
$> man 3 regex
Or google for regcomp() or regexec().
Giórgenes
2010-05-07 19:09:19