find number of lines with length more than 5

Question

Let's assume I've got a file

a
ab
abc
abcd
abcde
abcdef
abcdefg
abcdefgh
abcdefgh
abcdefg
abcdef
abcde
abcd
abc
ab
a

I need to find lines with length >= 5. Here is the code

int cOriginal(char *fileName)
{
    FILE *file = fopen(fileName, "r");
    char line[256];
    int numberOfStrings = 0;
    while(fgets(line, sizeof(line), file))
    {
        if (strlen(line) >= sizeCap)
        {
            numberOfStrings++;
        }
    }
    fclose(file);
    return numberOfStrings;
}

where sizeCap is a constant, defined as

const int sizeCap = 5;

My main() function

int main()
{
    char filename[] = "file.txt";
    int cnt3 = cOriginal(filename); // th
    printf("Using function3: %d", cnt3);
    return 0;
}

The problem is that it returns 10 instead of 8. And I don't get the idea why.

I don't have an IDE beside me so I have to use g++ and command line to compile it so I cannot debug it (or at least I don't know how).

What is the problem?

`fgets` keeps the new-lin character at the end of the line, which contributes to the `strlen`. You can `strtok(line, "\n")` to remove it. `strtok` is in `` — M Oehm, May 05 '15 at 08:34
Printing the lines that are longer than `sizeCap` would show you the problem. — Paul Hankin, May 05 '15 at 08:40
Compile with all warnings and debug info (`gcc -Wall -Wextra -g`). Then **use the debugger** (`gdb`); BTW, use `g++` for C++ code and `gcc` for C code, since C and C++ are different languages. — Basile Starynkevitch, May 05 '15 at 08:46

Sourav Ghosh · Accepted Answer · 2015-05-05T09:20:51.893

3

This is because fgets() reads and stores the trailing newline \n. This \n gets counted as one element while calculating strlen().

You need to strip off that \n after taking the input.

Manual Way
1. check the strlen().
2. check the n-1the byte for \n, replace with \0.
3. calculate the final(actual) strlen.
Using library function

using strchr()

You need to use strchr() to find out the trailing \n.
if found, replace the \n to \0 using the retuned pointer.

Also, you can get the new string length by subtracting the returned pointer from the original input string.

Sample Code:

int length = 0;
char*nl=strchr(line,'\n'); 
if (nl) 
{
    *nl='\0';
    length =  nl - line;
}
else
    length = strlen(line);

using strtok()

Have a look at the man page of strtok(). After reading the input, you can parse the string using \n as delimter.
calculate the strlen() of the returned token, if it is not NULL.

Some suggestions:

always check the return value of fopen() against NULL to ensure the success.
The recommended signature of main() is int main(void).

edited May 05 '15 at 09:20

answered May 05 '15 at 08:35

Sourav Ghosh

133,132
16
183
261

Why is it doing it? I mean add `\n` to the end of line – lapots May 05 '15 at 08:39
2

@user1432980 that is how `fgets()` behave. Check the link I provided for more details. – Sourav Ghosh May 05 '15 at 08:40
For finding the `\n`, `strchr` is much better -and faster- than `strtok`; and no need to compute `strlen` twice. – Basile Starynkevitch May 05 '15 at 08:43
@BasileStarynkevitch sir, 1) for _finding_ the presence of `\n` `strchr()` is better of course, but here, we need to strip off the `\n`, i.e., we need a _string_ without the `\n`, so in that case, is not `strtok()` a better choice? 2) You are right, that was a typo. Removed the double-`strlen()` check.. – Sourav Ghosh May 05 '15 at 08:56
1

No, just code `char*nl=strchr(line,'\n'); if (nl) *nl='\0';` (and you could even avoid calling `strlen` when an `\n` is so found, by using in that case `nl-line`) – Basile Starynkevitch May 05 '15 at 08:57
@BasileStarynkevitch I understood sir. Will update my answer accordingly. Thanks for the comment. :-) – Sourav Ghosh May 05 '15 at 09:01
I am surprized so many people are using `strtok`. I find that function disgusting and used it only once (since 1985), then gave up using it. – Basile Starynkevitch May 05 '15 at 09:02
@BasileStarynkevitch sir, _maybe_ the reason behind this is the ability to handle multiple delimiter at a time. Don't really know. :-) – Sourav Ghosh May 05 '15 at 09:08
Your revised code would crash when no `\n` is found – Basile Starynkevitch May 05 '15 at 09:09
@BasileStarynkevitch Thank you once again sir. Somehow missed that part. Updated accordingly. :-) – Sourav Ghosh May 05 '15 at 09:13
Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/76975/discussion-between-basile-starynkevitch-and-sourav-ghosh). – Basile Starynkevitch May 05 '15 at 09:17

find number of lines with length more than 5

1 Answers1