Reading text file into char array

Question

I'm having some trouble getting text into a char array. It worked fine when I set a static size for the array like

char speech[15000];

but that was inefficient so I tried using calloc instead. That made it stop working. The array exists at the right size, but nothing gets written in. Here's the relevant code. What am I doing wrong?

int main() {

FILE* inFile;
int i;
int count = 0;

printf("\nOpening file April_30_1789.txt\n");

inFile = fopen("./speeches/April_30_1789.txt", "r");

if(inFile == NULL) {
    printf("Could not find April_30_1789.txt\n");
    return -1;
}

char ch;

while((ch = fgetc(inFile) != EOF)) count++;

rewind(inFile);

int size = count;

printf("Size of the array is %d\n", size);

char *speech = (char *)malloc(size*sizeof(char) + 1*sizeof(char));

fscanf(inFile, "%s", speech);

printf("Closing the file.\n");
fclose(inFile);

printf("%s", speech);

printf("\n\nDone\n");

return 0;

}

Currently, this gives me

Opening file April_30_1789.txt
Size of the array is 8617
Closing the file.
Fellow-Citizens

Done

The *internal* file pointer is still pointing to the end of file. — pmg, Mar 27 '14 at 19:33
using fseek move the file pointer to the 0th offset and initialize the count to 0 in the begining — M Thotiger, Mar 27 '14 at 19:36
Why do you think static allocation is inefficient when compared to dynamic allocation? Have you performed any tests to prove that? Or are you afraid of wasting memory when you have less than 15000 bytes to read? This question is weird because static allocation incurs *less* overhead than dynamic allocation. If anything, you'd want dynamic allocation to let your buffer grow as needed. — Filipe Gonçalves, Mar 27 '14 at 19:38
@FilipeGonçalves Well, I have to use this program with text files of varying lengths, so I figured malloc/calloc would be better than a set number. — Gary M, Mar 27 '14 at 19:45
@user2797058 Fair enough. Turns out it wasn't about efficiency then. — Filipe Gonçalves, Mar 27 '14 at 19:46

score 10 · Accepted Answer · edited May 23 '17 at 12:17

10

Possible duplicate of Reading the whole text file into a char array in C.

Your problem : fscanf with the "%s" format will read up to the first whitespace encountered.

Possible solution (error-checking is omitted for conciseness) :

#include <stdio.h>  /* printf */
#include <stdlib.h> /* fopen, fseek, ... */

char *buffer = NULL;
size_t size = 0;

/* Open your_file in read-only mode */
FILE *fp = fopen("your_file_name", "r");

/* Get the buffer size */
fseek(fp, 0, SEEK_END); /* Go to end of file */
size = ftell(fp); /* How many bytes did we pass ? */

/* Set position of stream to the beginning */
rewind(fp);

/* Allocate the buffer (no need to initialize it with calloc) */
buffer = malloc((size + 1) * sizeof(*buffer)); /* size + 1 byte for the \0 */

/* Read the file into the buffer */
fread(buffer, size, 1, fp); /* Read 1 chunk of size bytes from fp into buffer */

/* NULL-terminate the buffer */
buffer[size] = '\0';

/* Print it ! */
printf("%s\n", buffer);

edited May 23 '17 at 12:17

Community

1
1

answered Mar 27 '14 at 19:55

Chnossos

9,971
4
28
40

Well, this worked out really well. It looks cleaner than what I had before (never used fseek/fread before) and I may just stick with this. For future reference though, what was my original code doing wrong? – Gary M Mar 27 '14 at 20:02
`fscanf` with the `"%s"` format will read up to the first whitespace encountered. For instance if I use your code to print my main.c file, only `#include` is actually printed. – Chnossos Mar 27 '14 at 20:10
Ah, that explains it. So if I had wanted to use my original code, I would just have to have it keep reading till EOF then, right? – Gary M Mar 27 '14 at 20:22
You could use a loop, but since whitespaces are discarded, you would end up loosing your document's formatting. Using `fread` will keep the formatting. Oh and by the way, this line `while((ch = fgetc(inFile) != EOF)) count++;` should be rewritten as `while((ch = fgetc(inFile)) != EOF) count++;`. – Chnossos Mar 27 '14 at 20:28

nikola-miljkovic · Answer 2 · 2014-03-27T20:03:24.347

1

as pmg said while((ch = fgetc(inFile) != EOF)) count++; makes your file pointer point at the end, use rewind(FILE* fileptr); to return to beginning of file.

edited Mar 27 '14 at 20:03

answered Mar 27 '14 at 19:37

nikola-miljkovic

652
4
14

This seems to work, but it only worked for the first word of the text, which was "Fellow-Citizens". Would that be a problem in reading the text into the array or is it an error with printing the array? – Gary M Mar 27 '14 at 19:44
@user2797058 Change `fscanf(inFile, "%s", &speech);` to `fscanf(inFile, "%s", speech);` – Filipe Gonçalves Mar 27 '14 at 19:47
@FilipeGonçalves Whoops, forgot to mention that I already did. I'll update the OP. It still only prints out Fellow-Citizens – Gary M Mar 27 '14 at 19:49
@nikolaMM94 It seems like the array is the correct size. Microsoft Word gave about the same number of characters that the variable count is reporting. If I did malloc right, it should be the right size – Gary M Mar 27 '14 at 19:54
@nikolaMM94 Would this work? char *speech = (char *)malloc(size*sizeof(char) + 1*sizeof(char)); This is what I have right now and it still gives only Fellow-Citizens – Gary M Mar 27 '14 at 19:57
Yes, no need 1*sizeof, sizeof and if that doesn't work try reading into buffer. – nikola-miljkovic Mar 27 '14 at 19:57

score 1 · Answer 3 · answered Mar 27 '14 at 19:40

Your file pointer inFile is pointing to end

I'd do something like following :

long lSize;

fseek( inFile , 0L , SEEK_END); //use the function instead
lSize = ftell( inFile );       // to know the file size
rewind( inFile );             // Now point to beginning 

char* speech = calloc( 1, lSize+1 );
if( speech )
{
    if( fread( speech , lSize, 1 , inFile) != 1)
    {
      fclose(inFile) ;
      free(speech); 
      exit(1);
    }
}

// Process the speech Here


fclose(inFile); 
free(speech); // Don't forget to free the allocated memory !

Reading text file into char array

3 Answers3