C - Get name from the string

[poziga]

Hello,

I have a question on how could I easily get name that is written on the beggining of a line in a file in C programming language not C++ or C#. Lets say I have a file test.txt and in it:

firsttext 2 4 2 2 2 -1 -99999
secondtext 3 3 13.4 3 3 3 3 3 3 3 5 5 10.2 7.3 -99999

I would like to get the name of a string in the beginning (firsttext or secondtext) of a line that has the most numbers that are 2% away from the value a user has written.

Lets say the user writes 2 and the program will return firsttext, if he writes 3 the program will return secondtext.

Thanks for your help in advance.

[kbw]

Read the whole line into a buffer (with fgets()).
Read the first string from the buffer (with sscanf()).

http://pubs.opengroup.org/onlinepubs/007908799/xsh/fgets.html
http://pubs.opengroup.org/onlinepubs/007908799/xsh/sscanf.html

[poziga]

Well if I do that i will only end up with the first word. I wont really solve the problem.

[Chervil]

You could use strtok to split the string into its component parts.
http://www.cplusplus.com/reference/cstring/strtok/

To convert a string to a number, the function atof or strtod might be useful.
http://www.cplusplus.com/reference/cstdlib/atof/
http://www.cplusplus.com/reference/cstdlib/strtod/

(since strtod can also set a pointer to the first character after the number, it might be possible to apply it repeatedly to get all the numbers in the string).

[willf]

I'm about as much of a newb when it comes to programming as is possible so take what I say with a grain of salt. But this can be solved with just math: write the function for standard deviation but substitute the input number in for the "average" in that equation. Calculate the deviation for each line and then pick the lower one.

[poziga]

Since i must read the text from file only once how would i save it to buffer and then call one line for example to get the first word and convert all the numbers into doubles? So i would have words seperated from numbers?

[Chervil]

You could read a single line at a time, using fgets. Then process that line. The result of that processing will be two items, the name, and count of values meeting the requirements. Store those in a char string and an integer count.

Then read and process the next line. If the count is greater than the stored value, replace the stored values with the current values.

I would advise using a separate function to handle the detailed processing of each line.

After processing all the lines, the stored values should contain the required result.

[poziga]

And how can i jump to new line after I have read a space char or -99999. I have the following code:

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

int main (int argc, char *argv[]){
char* besede[10];
char tabela [99];
FILE *file = fopen("test.txt", "r" );
if (file == 0){
printf("Napaka!\n");
}
else{
char x;
int i = 0;
int j = 0;

//while(x = fgetc(file) != EOF){
while((x = fgetc(file)) != ' '){
tabela[i] = x;
printf("%c", x);
i++;
}

besede[j] = tabela;
x = 0;
j++;
//}
fclose(file);
/*for(int k = 0; k <= 10; k++){
printf("%s",besede[k]);
}*/
}
}

[Chervil]

It would be easier based in your first post, to use fgets() to read a whole line at once, rather than reading a single character at a time.
http://www.cplusplus.com/reference/cstdio/fgets/

After that, split the string into words or 'tokens' using strtok()
http://www.cplusplus.com/reference/cstring/strtok/

Convert a character string into a number using atof()
http://www.cplusplus.com/reference/cstdlib/atof/

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

int main (int argc, char *argv[])
{
    char linebuf[1000];

    FILE *file = fopen("test.txt", "r" );

    if (file == 0)
    {
        printf("Napaka!\n");
        return 1;
    }

    while (fgets(linebuf, 1000, file))
    {
        /* printf("%s", linebuf); */

        char * pch;
        double num;
        printf ("Splitting string into tokens:\n");

        pch = strtok (linebuf, " ");
        if (pch != NULL)
        {
            printf("Name = %s\n", pch);
            pch = strtok (NULL, " ");
        }

        while (pch != NULL)
        {
            /* printf ("%s\n",pch); */

            num = atof(pch);
            printf("Number = %f\n", num);

            pch = strtok (NULL, " ");
        }

        printf("\n\n");
    }

    fclose(file);
    return 0;
}

Input:

firsttext 2 4 2 2 2 -1 -99999
secondtext 3 3 13.4 3 3 3 3 3 3 3 5 5 10.2 7.3 -99999


Output:

Splitting string into tokens:
Name = firsttext
Number = 2.000000
Number = 4.000000
Number = 2.000000
Number = 2.000000
Number = 2.000000
Number = -1.000000
Number = -99999.000000


Splitting string into tokens:
Name = secondtext
Number = 3.000000
Number = 3.000000
Number = 13.400000
Number = 3.000000
Number = 3.000000
Number = 3.000000
Number = 3.000000
Number = 3.000000
Number = 3.000000
Number = 3.000000
Number = 5.000000
Number = 5.000000
Number = 10.200000
Number = 7.300000
Number = -99999.000000

[poziga]

I did it a bit differently like this, but i have a problem why printf prints -99999 instead of firsttext and the same with secondtext. I already have the seconc part done I only have to figure out how to put the string into tabel if it is a word or if it is a number, then somehow transform char * stevila[1000] to double stevila[1000] :

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

int main (int argc, char * argv[]) {
    char tabela[99];
    
    char * besede[1000];
    char * stevila[1000];
    
    int stevecBesed = 0;
    int stevecStevil = 0;
    
    
    FILE *f = fopen("test.txt","r");
    
    while(fscanf(f, "%s", &tabela) != EOF){
        
        if(atof(tabela) == NULL){
            besede[stevecBesed] = tabela;
            printf("%s ", besede[stevecBesed]);
            stevecBesed++;
        }
        else{
            stevila[stevecStevil] = tabela;
            stevecStevil++;
        }        
    }
    //printf("%s", besede[0]);
    
    fclose(f);
    return 0;
}

[poziga]

If anyone knows how to fix this. I need it quickly. Thank you.

[Chervil]

The use of scanf seems a reasonable method.

But be careful when trying this if(atof(tabela) == NULL)
My compiler rejects this line as an error:

[Error] invalid operands to binary == (have 'double' and 'void *') 

Even if the compiler accepts it, if the input file contains any number which is zero, then atof(tabela) will result in zero, therefore this test cannot distinguish between a valid number "0" or some other text "abc".

i have a problem why printf prints -99999 instead of firsttext and the same with secondtext

Line 20 besede[stevecBesed] = tabela; is simply copying the address of tabela to the array. If the contents of tabela changes, so does the data pointed to by besede[stevecBesed].

That could be fixed like this, change the array definition to allocate space for the contents of the string
char besede[1000][20];

and use this to copy the string to the array:
strncpy(besede[stevecBesed], tabela, 20);


as for the numeric values,
double stevila[1000];

stevila[stevecStevil] = atof(tabela);

[poziga]

Thank you very mutch.

[poziga]

Oh one more thing. You mentioned the invalid operands to binary error. I didn't get it on my computer but got it on our schools online compiler. How do I fix that error?

[Chervil]

A simple fix for the compiler error is to use this, put 0.0 instead of NULL

    if (atof(tabela) == 0.0)

However, if any of the numbers can be zero, there is still a logic error.

You could try to test whether or not the string contains a number, by checking for '0' to '9' or '-' as the first character.
So instead of this:

    if (atof(tabela) == 0.0)

you could try this:

if (!(tabela[0] >= '0' && tabela[0] <= '9') && tabela[0] != '-')

... though I don't like that very much :)