我的fscanf为什么不读？

Question

I'm trying to read information from the file.txt into BST. First, I was thinking of saving info into temporary mini-arrays, and than assigning those values to the tree nodes. Apparently something goes wrong, I'm sure that one thing is the incorrect regular expression and might be something else. Please help me to figure out how to save data to the BST, I'm really late with this project.

Samples of data contained in the input file.txt :

3800 Lee, Victor; 2.8
3000 Brown, Joanne; 4.0
1300 South, Frankie; 2.6
4000 Trapp, Dave; 3.9

They have to be read and parsed into: 1) int= id , 2) char= name , 3) float= gpa

They have to be saved properly, without any symbols, so when I will do search in my next functions it wont give me any errors.

Here is my code below(UPDATED):

int readFile(BST_TREE *list) {
int id  = 0;
int val = 0;
int ln  = 0;
float gpa = 0.0;
char name[MAX_NAME_LEN];
STUDENT *stuptr = 0;
char line[MAX_LEN];

stuptr = (STUDENT*)malloc(sizeof(STUDENT));


FILE *fp = fopen("gpa.txt", "r");
if(fp == NULL) {
    printf("Error in opening file\n");
    return 1;
}

while (fgets(line, sizeof(line), fp)) {

ln++;
if (sscanf(line, "%d %[^;]; %f ", &id, name, &gpa) < 3) {
    printf("Syntax error in line %d.\n", ln);
} else {
    printf("1 %4d  %-29s%f\n", id, name, gpa);
  }
 }

    stuptr->id = id;
    strcpy(stuptr->name, name);
    (stuptr->gpa) = gpa;
    return 0;
}

// Typedef
 typedef struct
{
    int   id;
    char  name[MAX_NAME_LEN];
    float gpa;
} STUDENT;

Answer 1

That is because the format string of fscanf is wrong. What you need is "%d %[^;];%f " . This means

1. First read a decimal integer and store it.
2. Then read and discard any number of whitespace characters.
3. Then match any number of all characters except a semicolon ; and store it. The buffer where the characters are written must be large enough for the characters read plus a terminating null byte added by fscanf .
4. Then match a semicolon ; and discard it
5. Then read a floating-point value discarding any number of leading whitespace characters.
6. Then read and discard any number of whitespace characters.

As you see, if any line in the file is ill-formatted, then fscanf will fail due to matching failure and might leave the file pointer in an unknown location.

You should use fgets instead to read a line and then use sscanf to process the line and read elements from it. Here's how you should do it. However, you must know beforehand, the maximum length a line in the file can have.

There are other mistakes in your code snippet.

1. You are allocating memory to store a pointer, not the structure itself by this-
   stuPtr = (STUDENT*)malloc (sizeof (stuPtr));
   You should do the following. Also don't cast the result of malloc and check its result for NULL -
   stuPtr = malloc(sizeof *stuPtr);

2. You cannot assign arrays. You should copy each character of the array instead. Also, the structure member name is a character, not an array.
   stuPtr->name = tempName;
   In the above statement, tempName is an array which decays into a pointer to its first element. This pointer is assigned to structure member name which is a character - this is clearly an error. You should change the type of structure member name into a char array.

Here's a modified version with the changes I suggested.

#define MAX_LEN 100
#define MAX_NAME_LEN 40

typedef struct {
    int id;
    char name[MAX_NAME_LEN]; 
    float gpa;
} student;   

void add_student(BST_TREE *list) {
    int id, val;
    float gpa;
    char name[MAX_NAME_LEN]
    student *stuptr;
    char line[MAX_LEN];

    FILE *fp = fopen("gpa.txt", "r")       
    if(fp == NULL) {
        printf("error in opening file\n");
        return 1;
    }       

    while(fgets(line, sizeof line, fp) != NULL)
        stuptr = malloc(sizeof *stuptr);
        if(stuptr == NULL) {
            printf("not enough memory to allocate\n");
            return 1;
        }
        val = sscanf(line, "%d %[^;];%f", id, name, gpa)
        if(val != 3) {
            printf("input format error\n");
            // handle it
        }
        stuptr->id = id;
        strcpy(stuptr->name, name);
        stuptr->gpa = gpa;
    }
}

Answer 2

About reading your file: The return value is not necessarily EOF on failure to convert all placeholders and your format string does not correspond to the format in your file.

fscanf returns the number of successful conversions. When the input has run out, it returns EOF . In your case, the format string will convert only the first item, so the return value is 1. You are looking for 3. (When fscanf does not convert all items, it is stuck. When you try to read the next line, you really try to read from where fscanf gave up converting, which is usually not where you think it is.)

The format of the entries in file is:

3800 Lee, Victor; 2.8

Your format string is:

"%4d, %[^\t\n,;]c %f"

Several things here.

• There is no comma after the id, but your format string has one.
• The spurious c after the brackets cannot be understood
• You treat the comma as one of the characters that end the name, but all names contain commas.
• The name entry ends with a semicolon, so there should be a semicolon in your format string where the c is.
• Names that are longer than 30 characters will lead to buffer overflows. Better specify a maximum string length.

So your format should look like:

"%d %29[^;]; %f "

Putting it all togehter:

while (fscanf (f, "%d %[^;]; %f ", &id, name, &gpa) == 3) {
    printf("1 %4d  %-20s%f\n", id, name, gpa);
}

This will stop on the first line that can't be converted, so if you have a typo, a missing semicolon perhaps, in the first line, nothing will be read. That is not desirable. Because the format of your file is line-based, I recommend to read lines with fgets first, then parse the line with sscanf :

char buffer[80];
int ln = 0;

while (fgets(buffer, sizeof(buffer), f)) {
    int id;
    float gpa;
    char name[30];

    ln++;
    if (sscanf(buffer, "%d %[^;]; %f ", &id, name, &gpa) < 3) {
        printf("Syntax error in line %d.\n", ln);
    } else {
        printf("1 %4d  %-20s%f\n", id, name, gpa);
    }
}

Now bad lines are just skipped, with a (terse) error message.

Oh, and concentrate on the real problem and don't get too fancy. While cramming everything together like this:

if(!(fpStudents = fopen("gpa.txt", "r")))
    printf("-*- Could not open file for reading! -*-\n"), exit(100);

might give you h4xx0r credits, your code is rendered better as:

fpStudents = fopen("gpa.txt", "r");
if (fpStudents == NULL) {
    printf("-*- Could not open file for reading! -*-\n");
    exit(100);
}

You can see everything you need at a glace: Open a file. If that goes wrong, print a message and exit. I have a hard time to read that from your code without squinting. Assignments in conditionals and comma operators are so seventies .

Answer 3

My notes on your code:

sizeof (stuPtr) should be sizeof (STUDENT).

You have to pass the type definination to size of, the allocate the number of bytes of that type.

Here you can find more about that