strtok question

[Stu Cazzo]

Hi all,
I have a question on why strtok is doing what it's doing for my
splitString( string2 ); call.

Below is the output for the entire program:

token was: word1
token was: word2
token was: word3
token was: word1
token was: word3
empty field found - token <(null)>


The splitString( string1 ); works as expected, 3 tokens are found.
The splitString( string2 ); does not work as I expected.
I was expecting this:

token was: word1
empty field found - token <(null)>
token was: word3

Why does it not see the empty field for lineToken2?
Is there a better way to strip out the tokens for the case I have
where
there is no space between my delimiter?

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

void splitString( char *string )
{
const char lineDelimiter[] = ",";
char *lineToken1;
char *lineToken2;
char *lineToken3;

lineToken1 = strtok( string, lineDelimiter );
if ( lineToken1 == '\0' )
{
printf("empty field found - token <%s>\n", lineToken1);
}
else
{
printf("token was: %s\n", lineToken1);
}

lineToken2 = strtok( NULL, lineDelimiter );
if ( lineToken2 == '\0' )
{
printf("empty field found - token <%s>\n", lineToken2);
}
else
{
printf("token was: %s\n", lineToken2);
}

lineToken3 = strtok( NULL, lineDelimiter );
if ( lineToken3 == '\0' )
{
printf("empty field found - token <%s>\n", lineToken3);
}
else
{
printf("token was: %s\n", lineToken3);
}

}



int main (int argc, const char **argv)
{
char string1[] = "word1,word2,word3";
char string2[] = "word1,,word3";

splitString( string1 );
splitString( string2 );

return( 1 );

}

 

[Jens Thoms Toerring]

Hi all,
I have a question on why strtok is doing what it's doing for my
splitString( string2 ); call.

Below is the output for the entire program:

token was: word1
token was: word2
token was: word3
token was: word1
token was: word3
empty field found - token <(null)>

The splitString( string1 ); works as expected, 3 tokens are found.
The splitString( string2 ); does not work as I expected.
I was expecting this:

token was: word1
empty field found - token <(null)>
token was: word3

Why does it not see the empty field for lineToken2?

Because that's not how strtok() works. The man page on my machine
for strtok() actually makes it rather clear:

A sequence of two or more contiguous delimiter characters in the
parsed string is considered to be a single delimiter. Delimiter
characters at the start or end of the string are ignored. Put
another way: the tokens returned by strtok() are always non-empty
strings.

So if you have more than one ',' in a row all of them are treated
the same as a single ','.

Is there a better way to strip out the tokens for the case I have
where there is no space between my delimiter?

I guess you will have to write your own function for that, probably
repeatedly using strchr() or strstr().

Regards, Jens

 

[Ben Bacarisse]

while (*src && (tknchar != *src)) {

}
if (*src && (tknchar == *src)) src++;

Some people might find that test confusing. It is certainly
belt-and-braces code.

 

[Antoninus Twink]

Some people might find that test confusing. It is certainly
belt-and-braces code.

Most people grow out of this sort of thing within a few months or so of
their initial child-like excitement at discovering a language with so
many side effects. Clarity and ease of debugging become more valuable
than a transient smug feeling of cleverness.

Actually I'm pretty tolerant of different people's ways of laying out
code, indenting and the rest, but CBF really does seem to have total
anti-taste when it comes to code formatting. Perhaps the most irritating
thing of all is

This seems to me to be about as helpful as the infamous

i++; /* increment i by one */

 

[Richard]

Most people grow out of this sort of thing within a few months or so of
their initial child-like excitement at discovering a language with so
many side effects. Clarity and ease of debugging become more valuable
than a transient smug feeling of cleverness.

I have to agree 100% with this statement. Some people seem to take
pleasure in obfuscating their code.

Actually I'm pretty tolerant of different people's ways of laying out
code, indenting and the rest, but CBF really does seem to have total
anti-taste when it comes to code formatting. Perhaps the most irritating
thing of all is


This seems to me to be about as helpful as the infamous

i++; /* increment i by one */

Again agreed.

 

[Ben Bacarisse]

You should have left the body.

It has no bearing on my point. Any loop body that terminates in the
normal way would do just as well. Maybe I should have said "<snip
body with no break statement>".

That code doesn't have the same
effect. Assuming I am correctly interpreting your message.

I think you missed the point. If you want the function to work when
tknchar might be 0, then if (tknchar == *src) src++; is enough. If
you don't want it to work when tknchar is 0 (as seems to be the case)
then if (*src) src++; is enough.

It is not in any way wrong, just as if (c == '\n' && c) is not really
wrong -- it just makes the reader do an unwarranted double take.

 

[Ben Bacarisse]

Most people grow out of this sort of thing within a few months or so of
their initial child-like excitement at discovering a language with so
many side effects. Clarity and ease of debugging become more valuable
than a transient smug feeling of cleverness.

I don't know what you mean in this case. I agree with the sentiment,
but what has it do with this example? How do you write this without
using side effects?

The reference to debugging makes me thing you object to the layout.
Would moving the increment of src down a line make it all OK?

Actually I'm pretty tolerant of different people's ways of laying out
code,

Ah. So you were commenting on the layout. What was the reference to
abusing side effects about?