Fast double file read in C - c

Question

I have a large file containing floating point numbers and I want to read them.
52.881 49.779 21.641 37.230 23.417 7.506 120.190 1.240 79.167 82.397 126.502 47.377 112.583 124.590 103.339 5.821 24.566 38.916 42.576
This is just the beggining of the file. It has 10000000 numbers.
I got this code but I don't know how to print the numbers.
#include <stdio.h>
#include <stdlib.h>
#include <err.h>
#include <fcntl.h>
#include <sysexits.h>
#include <unistd.h>
int main()
{
int fd;
size_t bytes_read, bytes_expected = 1000000*sizeof(double);
double *data;
char *infile = "file.dat";
if ((fd = open(infile,O_RDONLY)) < 0)
err(EX_NOINPUT, "%s", infile);
if ((data = malloc(bytes_expected)) == NULL)
err(EX_OSERR, "data malloc");
bytes_read = read(fd, data, bytes_expected);
if (bytes_read != bytes_expected)
err(EX_DATAERR, "Read only %d of %d bytes",
bytes_read, bytes_expected);
/* print all */
free(data);
exit(EX_OK);
}

Answer 1

You are attempting to read a text file as if the data was binary, so you will read some bytes but the double values stored in the array will not be the values that you wanted to read from the file, you can probably do this
FILE *file;
double *array;
size_t count;
const char *infile = "file.dat";
file = fopen(infile, "r");
if (file == NULL)
return -1;
count = 0;
while (fscanf(file, "%*lf") == 1)
count += 1;
rewind(file);
array = malloc(count * sizeof(*array));
if (array == NULL) {
fprintf(stderr, "cannot allocate %zu bytes!\n", count * sizeof(*array));
fclose(file);
return -1;
}
// Read the values into the array
for (size_t i = 0; i < count; ++i) {
fscanf(file, "%lf", &array[i]);
}
// Print the array
for (size_t i = 0; i < count; ++i) {
fprintf(stdout, "%f\n", array[i]);
}
// Release memory
free(array);

Answer 2

Since you want a fast solution, maybe you have to sacrifice memory.
The faster manner of reading a file is in binary form.
Thus, I would obtain the file size with an efficient method,
then I would allocate memory accordingly,
with the idea of uploading the entire file to memory.
There, since memory reading is faster than file reading,
the data can be quickly read by using sscanf(...).
We can also observe that each floating point number
needs at least 3 characters to be stored in a text file:
1 char for the dot ('.'),
1 char for some digit,
and 1 char for
a space (' ') used to separating a value from its succesor in the
file.
Thus, the file size divided by 3 will be the upper bound for the size of the array of doubles.
#include <stdio.h>
int main(void) {
char *filename = "file.dat";
FILE *F = fopen(filename, "rb");
fseek(F, 0L, SEEK_END);
long int filesize = ftell(F);
rewind(F);
char *data = malloc(filesize+1);
fread(data, filesize, 1, F);
data[filesize] = '\0'; // End of string, just in case
fclose(F);
// The desired data will be stored in array:
double *array = malloc(sizeof(double) * filesize/3);
int ret;
int n; // represents the no chars in a sscanf(...) reading
double *a = array;
while (1) { // Infinite loop...
ret = sscanf(data, " %lg%n", a, &n);
if (ret == EOF) break; // <<---- EXIT POINT of the loop
a++;
data += n;
}
long int array_size = a - array + 1;
}