How can get int - string - float from file? - c

Question

I have a file like this one:
1234 Homer 18.5
1223 Bart 25.5
9341 Lisa 30.0
3420 Marge 28.4
8730 Abram 26.7
1876 Barns 27.8
1342 Smiters 23.0
7654 Milhouse 29.7
How can i get the first part ( for example 1234 ) of each line?
And how can i get the name ( for example Homer ) of each line?
I wrote this code below:
#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
int main()
{
char ch[25];
int i, num;
FILE *fp;
fp = fopen("studenti.txt","r"); // read mode
if( fp == NULL )
{
perror("Error while opening the file.\n");
exit(EXIT_FAILURE);
}
printf("The contents of numeri.txt file are :\n");
for(i = 0; i < 25; i++){
while( ( ch[i] = fgetc(fp) ) != EOF ){
if(!(ch[i] >= 'A' && ch[i] <= 'Z') && !( ch[i] >= 'a' && ch[i] <='z')){printf("%c",ch[i]);}
}}
fclose(fp);
return 0;
}
How can do that??

Answer 1

This is what fscanf function is for:
int n;
char name[25];
float x;
FILE* fp = ...
while (fscanf(fp, "%d%24s%f", &n, name, &x) == 3) {
// Do something with the data you just read:
printf("int=%d name='%s' float=%f\n", n, name, x);
}
Several things to note about the above:
fscanf returns the number of items it read from the file. Continue calling fscanf while it returns 3
%24s means "a string of up to 24 characters in length". name has 25 characters, because the last one is used for null termination
int and float parameters are passed to fscanf with an ampersand, because fscanf needs a pointer. String, on the other hand, takes no ampersand, because it's equivalent to a pointer.

Answer 2

If you are sure of the text format, the simplest may be to use fscanf.
int num;
char name[1024];
float grade;
fscanf(fp, "%d %s %f", &num, name, &grade);
Be aware, if the name is longuer than 1024 chars, you will have buffer overflows. If the format is not sure, you need to check the return code of fscanf (see the man page).

Answer 3

The only thing you ever read from files is bytes.
The first step is to check if the bytes are valid characters, and convert the bytes into characters if necessary. This is not necessarily simple. If the bytes are supposed to be ASCII, then you might only need to check if the bytes are valid ASCII (e.g. not less than or equal to zero and not above 0x80; and possibly not control characters like "delete" or "vertical tab").
However, where names are involved it's extremely unlikely that ASCII is adequate. This means you want something like UTF-8. In that case, at a minimum you need to check if the bytes are valid (variable length) UTF-8 sequences; in addition to checking for invalid characters (like "delete" or "vertical tab").
More complicated is if you simply don't know what the bytes are. There are ways to auto-detect the character encoding (but it's heuristics not 100% reliable).
The second step is parsing. Parsing typically has 2 equally important goals. The first goal is to convert the characters into a more easily processed form - e.g. like maybe a structure with 3 fields (an integer, string and float) representing each line of characters. The second goal of parsing is reporting any errors to the user in an easily understood manner.
For example, maybe the first number on each line must be a 4 digit code (like "0123"); and if there's only 3 digits (like "123") then you want to generate an error (e.g. "ERROR: CourseID too short on line 5 of file 'foo.txt'") so that it's easy for the user to know exactly what the problem is and easy for the user to fix it.
Note: I don't think I've ever seen code that uses fscanf() that is close to (what I consider) acceptable. There's almost never useful/descriptive error messages.