Thread: Confused parsing a CSV txt file... strtok, fgets, sscanf

Hi everyone.

I'm working with a comma-delimited text file (in this example, it's abc.txt), with the following structure:

Code:

299, 26, 10, 7, 45, 87.688493, 112.055298, 54697, 3362, 1992, 2569, 793, 81638, 3
299, 26, 10, 8, 0, 85.539068, 114.966376, 49899, 5755, 3751, 3014, 994, 89688, 5
299, 26, 10, 8, 15, 83.402953, 117.925606, 47183, 7941, 4041, 3164, 1084, 93264, 5

I'd like to parse this code into fourteen separate arrays, so each 'column' is in its own array. For example, the output for the first four items of each line should be in this structure:

Code:

array1[299,299,299]
array2[26,26,26]
array3[10,10,10]
array4[7,8,8]

I'm incorporating this code into a program that can read from different user-selected text files which have different numbers of lines, but will always the same fourteen-variable structure. From what I've gathered, I'll need to use a While loop with fgets inside of it to accomplish this.

I've tried several different methods to get the data into the arrays, such as strtok, sccanf, and a combination of the two. Everything I've ended up with was a bunch of garbage numbers, or blank space.

Please forgive any noobie mistakes I may have made. This is my first outing with C programming. Here's the code I've worked with so far:


This first attempt I made at the code produces garbage numbers.

Code:

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

int main()
{

char line[200];
int i;
int aJulianDay[50];
int aDay[50];
int aMonth[50];
int aHour[50];
int aMinute[50];
float a1[50];
float a2[50];
int a3[50];
int a4[50];
int a5[50];
int a6[50];
int a7[50];
int a8[50];
int aHazardRating[50];


	FILE *pFile;
	pFile = fopen("c:\abc.txt", "rt");



// Load each different 'column' of values into an array


//fgets gets a string from a file
//one string per line

    while(fgets(line, sizeof line, pFile) != NULL)
    {
// sscanf reads the data from the strings and stores it in arrays
// since the line is in a string, it finds each of the values and stores
// them into the specified arrays. The loop does this for each line
// so, each julian day etc will be read into array aJulianDay[] each time there's a line
// Thus each variable is stored into a different array 
    if(sscanf(line, "%d,%d,%d,%d,%d,%lf,%lf,%d,%d,%d,%d,%d,%d,%d", &aJulianDay[i], &aDay[i], &aMonth[i], &aHour[i], &aMinute[i], &a1[i], &a2[i], &a3[i], &a4[i], &a5[i], &a6[i], &a7[i], &a8[i], &aHazardRating[i]) == 14)
    {
    ++i;
    }
    }

//Close the file
fclose(pFile);

//Check the 2nd element of each array to see if it worked
printf("%d\n",aDay[1]);
printf("%d\n",aMonth[1]);
printf("%d\n",aHour[1]);
printf("%d\n",aMinute[1]);
printf("%d\n",a1[1]);
printf("%d\n",a2[1]);
printf("%d\n",a3[1]);
printf("%d\n",a4[1]);
printf("%d\n",a5[1]);
printf("%d\n",a6[1]);
printf("%d\n",a7[1]);
printf("%d\n",a8[1]);
printf("%d\n",aHazardRating[1]);

getch();

  return 0;
  
}

I scrapped that and took a different approach, using strtok to try to parse the files using the comma delimiters. I got as far as this:

Code:

//Working copy

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

int main()
{


	FILE *pFile;
	pFile = fopen("c:\abc.txt", "r");


//fgets gets a string from a file
//one string per line

    fgets(line, sizeof line, pFile);
	
	char delimiter[] = ",";
	char *token;
	int i5 = 0;
	int input[15];
	
	token = strtok(line, delimiter);
	while (token != NULL)
    
	{
	

    sscanf(token, "%d", &JulianDay1); 
    input[i5++] = JulianDay1[i5];
    
	token = strtok(NULL, delimiter);
	}
//Close the file
fclose(pFile);

//Test to see if elements are correct
printf("%d\n",JulianDay1[0]);
printf("%d\n",JulianDay1[1]);
printf("%d\n",JulianDay1[2]);
printf("%d\n",JulianDay1[3]);
printf("%d\n",JulianDay1[5]);
printf("%d\n",JulianDay1[7]);
printf("%d\n",JulianDay1[8]);
printf("%d\n",JulianDay1[9]);
printf("%d\n",JulianDay1[10]);
printf("%d\n",JulianDay1[11]);
printf("%d\n",JulianDay1[12]);
printf("%d\n",JulianDay1[13]);
printf("%d\n",JulianDay1[13]);

getch();



	  getch(); //See exactly what you're doing
  
  return 0;
  
}

But, that didn't work either. So, I tried another approach:

Code:

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

int main()
{
  

	char line[300];
	char delimiter[] = ",";
	char *token;
	int var;
	int input[15];
	int i = 0;

	FILE *pFile;
	pFile = fopen("c:\abc.txt", "r");  //Open that file 	in Read mode
	
    while(fgets(line, sizeof line, pFile)) //Read the file one line at a time until the EOF
	{
	
		token = strtok (line, delimiter); //Pull the string apart into tokens using the commas
		while (token != NULL)
			{
			sscanf (token, "%d", &var); //Scan that token into your placeholder
			input[i++] = var;
			printf("%d/n", input[i]); //Print out the array

			token = strtok (NULL, delimiter);
			}

	}
	
	fclose(pFile); //Close that file
	
	  getch(); //See exactly what you're doing
  
  return 0;
  
}

That didn't work either.

My head is spinning at this point, and I'm not sure which one of these methods I'm closest to getting any kind of valid output from. I would really appreciate any help in getting these fourteen arrays populated with my data. I've tried every avenue I've found, but I can't quite figure it out. Sorry in advance for any newbie mistakes I may have made.

Nice catch Quzah...

fscanf() and/or sscanf(), is probably the easiest input method for a beginner, because scanf() is the first input method generally taught in C. Strtok() involves using a pointer, so I doubt you'll like it.

so:

set i = 0, before the while loop in the first example, and double up on the \ char in the file name: c:\\etc.

Don't know if it applies to what you're doing here, but generally, I'd put 14 bits of data that were related, into one struct with 14 struct members (fields), and then make 1 array of those structs.

Originally Posted by stahta01

Code:

printf("%f\n",a1[2]);

Use %f to print an double.

Hmm, I changed the code as directed, and the output is still wrong....

Code:

printf("%f\n",a1[2]);
printf("%f\n",a2[2]);

on the same line as before outputs these:

Code:

-116933779947120400000000000000000.000000
0.000000

Am I doing something wrong with the sscanf code?

Originally Posted by stahta01

Note: You need to learn what an array of structure data type is.
Is the number of columns in your data fixed?
If yes, then using an structure makes sense to me, also.

Tim S.

Yes, the number of columns is always fourteen. Only the number of lines will vary. I'm not sure what a structure data type array is. What advantages would that have?

Originally Posted by stahta01

Readability and maintainability will be much better using a structure, look-up the struct keyword.

I now see problem with sscanf; found problem "%lf" is for double data type, you are using float data type.

You need to add more error checking.
Like adding and else to your "if(sscanf(line," statement that says which line had the error on it.
Checking to make sure the file open right and message and exit if it fails to open.

Tim S.

Aha! changing %lf to %f in sscan() fixed it. Thanks! I'm glad to know that I had the right idea from the start, and it was just minor things keeping my output from being correct.

I'll look into using struct, but honestly, at this point I think I'd rather stick to the code I have, unless there was an extremely compelling argument as to why I need to use struct.

Now onto figuring out how to search through the arrays using pointers... If I have issues with that later on, should I recycle this thread, or should I start a new one?

Originally Posted by towely

If I have issues with that later on, should I recycle this thread, or should I start a new one?

If the Title still applies add on to this thread (and, if this thread is not over a month old).
If not, start new thread with link back to this thread.

Tim S.

Sounds good, thanks.