读取csv文件到c结尾

Question

So I have this code that reads a bunch of lines from a csv document. I know that initially the document has 16 rows, and that's why I designated int noRows = 16; in my main funtion.

void readBeer(int noRows) {
char *oneline, *token;
char oneproduct[256];
char delim[] = ",";
int x = 1;

FILE *fp; //open file
if ((fp = fopen("varor.csv", "r")) == NULL) //can the file be opened?
{
fprintf(stderr, "File varor.csv couldn't be opened\n"); //"couldn't open file"
exit(-1);
}

while(noRows != 0) 
{
    int countTok = 1;

    fgets(oneproduct, 256, fp); //get the first row
    oneproduct[strlen(oneproduct) - 1] = '\0'; // remove end-of-line character

    oneline = strdup(oneproduct); //duplicate oneproduct into oneline because strtok modifies the given string
    token = strtok(oneline, delim); //split oneline into tokens, tokens are separated by ","

    while (token != NULL) 
    {
        if(countTok == 1) beer[x].productNumber = atoi(token);
        else if(countTok == 2) strcpy(beer[x].name, token);
        else if(countTok == 3) beer[x].price = atof(token);
        else if(countTok == 4) beer[x].volume = atof(token);
        else if(countTok == 5) strcpy(beer[x].type, token);
        else if(countTok == 6) strcpy(beer[x].style, token);
        else if(countTok == 7) strcpy(beer[x].packaging, token);
        else if(countTok == 8) strcpy(beer[x].country, token);
        else if(countTok == 9) strcpy(beer[x].manufacturer, token);
        else if(countTok == 10) beer[x].alcohol = atof(token);
        else printf("kossan hoppade!"); //should never be seen in console
        token = strtok(NULL, delim);
        countTok++;
    }

    x++;
    noRows--;
    free(oneline); free(token);

}

fclose(fp);

}

My question is how do I read the file to its end without first knowing how many rows it has? I'm thinking of having a specific cell in the file just to save noRows between startup and shutdown of console.

I tried using char buffer[1000]; while(fgets(buffer, 1000, fp)) {} char buffer[1000]; while(fgets(buffer, 1000, fp)) {} but then it reads the first 8 rows(not sure if it's always exactly 8) as 0,0,0,0,0,0,0,0,0,0.

Answer 1

The answer to your question if just test the return value of fgets

for (;;)   // idiomatic C style for an infinite loop 
{
    int countTok = 1;

    if (NULL == fgets(oneproduct, 256, fp)) break; //get one row and exit loop on EOF
    ...

But as you were told in comments, there are many other problems in your code:

• you initialize your indexes to 1 while C array indexes start at 0. So I assume that x=1; should be x=0; . For the same reason, countTok = 1; is not wrong, but countTok = 0; would be more idiomatic C
• you use fprintf(stderr, ... to notice an error when opening the file. While not wrong, it gives no indication on the cause of the error. perror would do...

• you erase last character of oneproduct without controling it is a newline. The assumption will be wrong

  • if one line contains at least 256 characters
  • if last line does not end with a newline

  The idiomatic way is to use strcspn :

   oneproduct[strcspn(oneproduct, "\\n")] = '\\0'; // erase an optional end of line 

• you duplicate oneproduct to oneline . Here again, nothing is wrong, but it is useless because you never use the original line
• the long list of else if could be replaced with a switch, but this one is mainly a matter of style
• you free token at the end of loop. As it is NULL it is a no-op, but it is was not allocated it should not be freed.

Answer 2

If you're using fgets , you can test if the result of the call is equal to NULL or not. fgets signals NULL on error or EOF (end-of-file).

char *fgets(char *line, int maxline, FILE *fp); .

IE:

while (fgets(buffer, MAXLINE, fp) != NULL) {
    // Process line here.
}

You can also process the entire file character by character and test if (c == EOF) . There are a couple of ways to do it.