C++ : read csv file with fgetc and separate words on semicolon ";" - c++

Question

I have to read in a csv file with 5 fields (int , char[], char[], char[], float) that looks like that :
2345678;Meier;Hans;12.10.1985;2.4;
1234567;Müller;Fritz;17.05.1990;1.9;
I have to put the fields in a struct, and then put the struct after one line is complete, into a array of the struct type ...
for the learning effect, we are only allowed to use LOW-LEVEL coding, and only use functions like fgetc, strcpy and no strings, only char[]...
Now I made my algorithm to read the textfile character by character, but I have problems separating them correctly, putting them together again and assigning them to the struct fields correctly. Here is my Code:
#include <cstdlib>
#include <stdio.h>
#include <stdlib.h>
#include <iostream>
#include <string.h>
using namespace std;
int main(int argc, char **argv)
{
struct Stud{
long matrnr;
char vorname[30];
char name[30];
char datum[30];
float note;
};
const int MAX = 30;
Stud stud;
Stud mystud[30]; // <<-- Array of "Stud" type
//memset((void*)mystud,0,sizeof(mystud) * sizeof(Stud));
int wordCounter(0);
int i(0); //thats the charCounter or index
int studentCounter(0);
char wort[MAX];
//int matrnr;
//char vorname[MAX];
//char name[MAX];
//char datum[MAX];
//float note;
FILE * pFile;
int cnr(0);
pFile=fopen("studentendaten.txt","r");
if (pFile==nullptr)
{
perror ("Fehler beim öffnen der Datei");
}
else
{
while (cnr != EOF)
{
(cnr=fgetc(pFile)) ;
if ((char)cnr == '\n') {
mystud[studentCounter] = stud;
studentCounter++;
continue;
}
if ((char)cnr == ';') {
wort[i] = '\0';
switch (wordCounter % 5) {
case 0:
stud.matrnr = atol(wort);
break;
case 1:
strcpy(stud.name, wort);
break;
case 2:
strcpy(stud.vorname, wort);
break;
case 3:
strcpy(stud.datum,wort);
break;
case 4:
stud.note = atof(wort);
break;
}
wordCounter++;
i = 0;
continue;
}
if (wordCounter % 5 == 0 && (char)cnr != ';') {
wort[i] = (char)cnr;
i++;
//stud.matrnr = atol(wort);
}
if (wordCounter % 5 == 1) {
wort[i] = (char)cnr;
i++;
//strcpy(stud.name, wort);
}
if (wordCounter % 5 == 2) {
wort[i] = (char)cnr;
i++;
//strcpy(stud.vorname, wort);
}
if (wordCounter % 5 == 3) {
wort[i] = (char)cnr;
i++;
//strcpy(stud.datum,wort);
}
if (wordCounter % 5 == 4) {
wort[i] = (char)cnr;
i++;
//stud.note = atof(wort);
}
}
fclose (pFile);
}
for (int i(0) ; i <= studentCounter; i++) {
cout <<mystud[i].matrnr << " " << mystud[i].name << " " << mystud[i].vorname <<" "
<< mystud[i].datum <<" " << mystud[i].note << endl;
//printf("%5ld %5s %5s %5s %5f \n",mystud[i].matrnr,mystud[i].name,mystud[i].vorname,mystud[i].datum,mystud[i].note);
}
return 0;
}
I am not sure if it has to do with a wrong increment variables, or the fact that I don't put an '\0' at the end of my wort[] array..and therefore not recognizing the end of my array? And if so, how do I do it without knowing where the end exactly is... ? (I don't know the length of the words..)
EDIT: I updated my code again, the only thing that wonders me is that the LAST LINE IS NOT BEING CORRECTLY PARSED , its showing some rubbish, and I can't see the error in my code...
2345678;Meier;Hans;12.10.1985;2.4;
1234567;Müller;Fritz;17.05.1990;1.9;
8392019;Thomas;Kretschmer;28.3.1920;2.5;
3471144;Mensch;Arbeit;29.2.2013;4.5;
2039482;Test;Test;30.20.2031;2.0;
7584932;Bau;Maschine;02.02.2010;2.3;
2345678;Meier;Hans;12.10.1985;2.4;
1234567;Müller;Fritz;17.05.1990;1.9;
8392019;Thomas;Kretschmer;28.3.1920;2.5;
3471144;Mensch;Arbeit;29.2.2013;4.5;
2039482;Test;Test;30.20.2031;2.0;
7584932;Bau;Maschine;02.02.2010;2.3;
2345678;Meier;Hans;12.10.1985;2.4;
1234567;Müller;Fritz;17.05.1990;1.9;
8392019;Thomas;Kretschmer;28.3.1920;2.5;
3471144;Mensch;Arbeit;29.2.2013;4.5;
2039482;Test;Test;30.20.2031;2.0;
7584932;Bau;Maschine;02.02.2010;2.3;
2345678;Meier;Hans;12.10.1985;2.4;
1234567;Müller;Fritz;17.05.1990;1.9;
8392019;Thomas;Kretschmer;28.3.1920;2.5;
3471144;Mensch;Arbeit;29.2.2013;4.5;
2039482;Test;Test;30.20.2031;2.0;
7584932;Bau;Maschine;02.02.2010;2.3;

Answer 1

Suggestion: use a case structure for the parsing, and make yourself a "copyToSemicolon" function: then you can write things like
sIndexCount = 0;
char temp[50];
while((cnr=fgetc(pFile)) != EOF) {
offset = 0;
for(var = 0; var < 5; var++ {
switch(var) {
case 0:
offset = copyToSemicolon(temp, cnr, offset) + 1;
stud.matrnr = atoi(temp);
break;
case 1:
offset = copyToSemicolon(mystud[sIndexCount].vorname, cnr, offset) + 1;
break;
... etc
}
}
sIndexCount++;
if(sIndexCount == 50) break; // in case the input file is longer than our structure
}
And you need a function copyToSemicolon that takes two char* pointers as inputs, and that copies characters from the second string (starting at offset) until it reaches either a semicolon or the end of line - and that returns the offset it reached (last character read).
int copyToSemicolon(char* dest, char* source, int offset) {
while(source[offset] != ';' && source[offset] != '\n') {
*dest = source[offset++];
dest++;
}
return offset;
}
EDIT strtok method:
sIndexCount = 0;
char temp[50];
while((cnr=fgetc(pFile)) != EOF) {
offset = 0;
temp = strtok(cnr, ';');
for(var = 0; var < 5; var++ {
switch(var) {
case 0:
stud.matrnr = atoi(temp);
break;
case 1:
strcpy(mystud[sIndexCount].vorname, strtok(NULL, ';'));
break;
... etc
case 4:
mystud[sIndexCount].note = atof(strtok(NULL, '\n'));
}
}
sIndexCount++;
if(sIndexCount == 50) break; // in case the input file is longer than our structure
}

Answer 2

One issue that I am seeing is that your code copies or parses one character at a time, such that when you're reading 2345678;Meier;Hans;12.10.1985;2.4; you first set stud.matrnr to 2, then 23, then 234, then 2345, then 23456, then 234567, then 2345678. Similarly, for stud.name, you first set it to M, then the Me, then to Mei, etc. I propose to you to think of things in a different way. I'll give you some pseudocode:
while (!eof) {
get character from file
if (character isn't ';' and isn't '\n') {
copy character into buffer (increment buffer index)
} else if (character is ';') {
it's the end of a word. Put it in its place - turn it to an int, copy it, whatever
reset the buffer
} else if (character is '\n') {
it's the end of the last word, and the end of the line. Handle the last word
reset the buffer
copy the structure
}
}
This should make life a lot easier on you. You're not changing your data nearly as much, and if you need to debug, you can focus on each part on its own.
Generally, in programming, the first step is making sure you can say in your native speaking language what you want to do, then it's easier to translate it to code. You're close with you implementation, and you can make it work. Just be sure you can explain what should be happening when you see ';' or '\n'.

Answer 3

Since you have tagged this as C++, you should consider using std::getline for reading the line from the file, the use std::getline(file, text_before_semicolon, ';') for parsing the fields.
You could also use std::istringstream for converting the textual representation in the text line to internal numeric format.