读取包含float的C中的.txt文本文件，用空格分隔

Question

it looks common and obvious but I've already read txt file in C in the past and here I am really stuck.

I have a txt file of this format

0.00000587898458 0.0014451541000 0.000000000001245
0.00012454712235 0.1245465756945 0.012454712115140

... with 640 columns and 480 lines.

I want to put each number of my txt file in a float with the maximum of precision as it is possible, and in a for loop.

FILE* myfile=NULL;
double myvariable=0.0;
myfile=fopen("myfile.txt","r");

for(i =0, k=0 ; i< height; i++)
    for (j=0 ; j< width ; j++){
fscanf(myfile,"%0.20f",&myvariable);
printf("%0.20f",myvariable);
k++;
}
}
fclose(myfile);

Thank you very much for your help

Answer 1

There are several errors in your program - mismatched braces, undefined variables, etc. The most important, however, and the one most likely to be causing your problem, is that you're not passing a pointer to myvariable in your fscanf() call. You'll want to use &myvariable there, so that fscanf() can fill it in appropriately. You probably don't need the format string to be so complicated, either - "%lf" should work just fine to read a double . In fact, gcc warns me with your format string:

 example.c:16: warning: zero width in scanf format example.c:16: warning: unknown conversion type character '.' in format 

And then my output becomes just 0. Try "%lf" . Here's a complete working example with your sample input:

#include <stdio.h>

#define HEIGHT 2
#define WIDTH  3

int main(void)
{
  FILE *myfile;
  double myvariable;
  int i;
  int j;

  myfile=fopen("myfile.txt", "r");

  for(i = 0; i < HEIGHT; i++)
  {
    for (j = 0 ; j < WIDTH; j++)
    {
      fscanf(myfile,"%lf",&myvariable);
      printf("%.15f ",myvariable);
    }
    printf("\n");
  }

  fclose(myfile);
}

Example run:

$ ./example 
0.000005878984580 0.001445154100000 0.000000000001245 
0.000124547122350 0.124546575694500 0.012454712115140 

Answer 2

 fscanf(myfile,"%0.20f",myvariable);

You have to pass a pointer, use &myvariable instead. Fix:

 fscanf(myfile, "%lf", &myvariable);

Answer 3

I can only guess, since you haven't shown us your actual source code, but ...

fscanf 's "%f" format expects a pointer to float ; you're giving it a pointer to double . That's probably what's causing the problem you're seeing. Use "%lf" for double (or "%Lf" for long double ).

Yes, this is an inconsistency between the *printf and *scanf functions. It exists because float arguments to *printf are promoted to double ; there's no such promotion for pointers, such as the float* arguments you might pass to fscanf .

And just use "%lf" , not "%0.20lf" ; I'm not sure the latter is even valid.

Answer 4

1. You say you are stuck, but do not get into specifics regarding what exactly is the problem you are encountering.
2. You most likely intended to pass the address of 'myvariable' to fscanf.
3. Even if you don't care so much about precision (and in your case, you clearly say tjat you do), you should prefer double over float. You should choose float only when you have a very strong reason for preferring float over double, and even that, should be in very specific situations.
4. Always check calls to fopen to ensure that it succeeded.