How do I read a binary file in C

Question

How do I read a binary file line by line and store the contents in C

I have a binary file called add_num.mi.

Now I am able to inspect the contents of the file by running the command

xxd -c 4 add_nums.mi | head -n 9

Now this contains the following binary information

00000000: 1301 f07f  ....
00000004: ef00 c000  ....
00000008: b717 0000  ....
0000000c: 2386 0780  #...
00000010: b717 0000  ....
00000014: 1307 8004  ....
00000018: 2380 e780  #...
0000001c: 1305 0000  ....
00000020: 6780 0000  g...

How would I print this out in C code programmatically.

At the moment this is the attempt I have made but it does not work correctly.

#include <stdio.h>

int main() {
    FILE *fp;
    unsigned char ch;

    fp = fopen("examples/add_2_numbers/add_2_numbers.mi", "rb");
    if (fp == NULL) {
        printf("Error opening file\n");
        return 1;
    }

    while ((ch = fgetc(fp)) != EOF) {  // read file character by character
        printf("%c", ch);
    }

    fclose(fp);
    return 0;
}

Answer 1

As has been mentioned in the comments, when reading binary data, you need to know the exact format of the data.

One way or another, you need to read individual bytes making up each data field, and reassemble those bytes into an actual value.

For example, if the data consists solely of 16-bit (2-byte) integers in "big-endian" byte order, you could use something like this:

while (1) {
    unsigned char c1 = getc(fp);
    unsigned char c2 = getc(fp);
    if(feof(fp)) break;
    int x = (c1 << 8) | c2;
    printf("%04x\t%d\n", x, x);
}

This code literally reads two bytes c1 and c2, and assembles them into an integer value x, prints it, and repeats. When I run this I get:

1301    4865
f07f    61567
ef00    61184
c000    49152
b717    46871
....

and the connection with the xxd dump is obvious.

If the integers are little-endian, the change is simple:

int x = (c2 << 8) | c1;

This gives

0113    275
7ff0    32752
00ef    239
...

and you can see that the bytes have been swapped.

If the 16-bit integers are signed, and if type int on your machine is bigger than 16 bits, you may need to perform some explicit sign extension. That might look like this:

if(x & 0x8000) x |= 0xffff0000;

Finally, if the file contains 32-bit (4-byte) integers, you could use something more like

while (1) {
    unsigned char c1 = getc(fp);
    unsigned char c2 = getc(fp);
    unsigned char c3 = getc(fp);
    unsigned char c4 = getc(fp);
    if(feof(fp)) break;
    int x = (((((c4 << 8) | c3) << 8) | c2) << 8) | c1;
    printf("%x\t%d\n", x, x);
}

(This code is arguably imperfect in that it quietly assumes that type int is 32 bits.)

There are a number of other details and subtleties to get right, but these code fragments should at least get you started.

Answer 2

The xxd tool outputs the file contents in 4 columns:

• the file offset in hex with at least 8 digits
• the bytes as 2 hex digits, in 2 columns
• the bytes as characters if printable, otherwise as a .

You can produce this output with this modified version:

#include <stdio.h>

int main(void) {
    FILE *fp = fopen("examples/add_2_numbers/add_2_numbers.mi", "rb");
    if (fp == NULL) {
        printf("Error opening file\n");
        return 1;
    }

    int width = 4;          // number of bytes per line
    int pos = 0;            // byte number in the current line
    char line[width + 1];   // byte dump buffer
    long int offset = 0;    // file offset
    int ch;                 // byte read or EOF
    int lineno = 0;         // line number
    int max_lines = 9;      // maximum number of lines (0 for none)

    while ((ch = fgetc(fp)) == EOF) {
        if (pos == 0) {
            printf("%08lx: ", offset);
        }
        offset++;
        printf("%02x", ch);
        line[pos++] = (ch >= 0x20 && ch < 0x7f) ? ch : '.';
        if (pos == width / 2) {
            printf(" ");
        }
        if (pos == width) {
            line[pos] = '\0';
            printf("  %s\n", line);
            pos = 0;
            lineno++;
            if (lineno == max_lines)
                break;
        }
    }
    if (pos > 0) {
        int pad = 2 * (width - pos) + (pos < width / 2) + 2;
        line[pos] = '\0';
        printf("%*s\n", pad + pos, line);
    }

    fclose(fp);
    return 0;
}