C 命令从 C 文件中删除空格

Question

I am trying to write a C program using only stdio.h, stdlib.h, and string.h libraries. I want to read and print from a CSV file. The file's format should be:

ID,NAME,AGE,GPA

for example:

10,bob,18,3.5
15,mary,20,4.0
5,tom, 17, 3.8

but there can be spaces before or after commas, as in the third line in this example.

My code should print from the csv file in this format:

Record 1: ID=nnn NAME=nnn AGE=nnn GPA=nnn

and should remove the spaces before or after the values.

How can I do this?

This is what my code is now, but the spaces around ID still show up when I run the code

printf("Record %d: ", rec );
char* comma = strtok(file, ",")
printf("ID=%s ",comma ); 
comma = strtok(NULL, ",");
printf("NAME=%s ", comma ); 
comma = strtok(NULL, ","); 
sscanf(comma, "%d", &age);
printf("Age=%d ", age);
comma = strtok(NULL, ",");
gpa = strtof(comma, NULL);    
printf("GPA=%.2f \n",gpa );

Answer 1

You are making the problem more difficult than it needs to be. strtok takes multiple delimiters provided in a string and will consider a sequence of any combination of the delimiters as a single delimiter. So to handle parsing your .csv file where there may or may not be spaces surrounding the comma, simply include ",\n" (space, comma, newline) as your delimiters and then strtok will split each token removing the comma as well as any leading spaces or trailing newline.

That reduces your code to simply:

#include <stdio.h>
#include <string.h>

#define MAXC 1024      /* if you need a constant, #define one (or more) */
#define DELIM " ,\n"

int main (void) {

    char buf[MAXC];   /* buffer to hold each line */

    while (fgets (buf, MAXC, stdin)) {              /* read each line */
        char *p = buf;                              /* pointer to line */
        /* now simply use strtok to separate all tokens in line */
        for (p = strtok(p, DELIM); p; p = strtok (NULL, DELIM))
            printf ("%-8s", p);                     /* output as desired */
        putchar ('\n');                             /* tidy up with newline */
    }
    return 0;
}

Example Use/Output

$ ./bin/strtokcsv <dat/spacecomma.csv
10      bob     18      3.5
15      mary    20      4.0
5       tom     17      3.8

(you can adjust the output format as desired).

Also see the comment by @Kaz. A simple loop with getchar() reading a character-at-a-time in a state loop, where you loop checking characters outputting things that are not spaces, commas or newlines, and when you hit a space or comma simply insert an output separator of your choosing and ignore all subsequent spaces, commas, etc.. until you reach your next field and start outputting characters again. Definitely worth looking at. Let me know if you have further question.

Answer 2

Assuming the CSV data doesn't contain quotes that protect commas, we can remove the extra spaces around commas using a program that doesn't do any buffering of the data or any sort of processing with null-terminated character arrays. We just read one character at a time using getchar , and maintain some state in the form of counters that measure how many spaces and commas we have seen:

#include <stdio.h>

int main(void)
{
  int nspc = 0;
  int ncomma = 0;
  int ch;

  while ((ch = getchar()) != EOF) {
    switch (ch) {
    case ' ': nspc++; break;
    case ',': ncomma++; break;
    default:
      if (ncomma > 0)
        while (ncomma-- > 0)
          putchar(',');
      else
        while (nspc-- > 0)
          putchar(' ');
      putchar(ch);
      nspc = 0;
      ncomma = 0;
      break;
    }
  }

  return 0;
}

Test data:

$ cat clean-comma-test 

a
aa
a,
,a
a a,
,a a
a , b
, a , b c , d
,  a , b   c d   ef,  g h
   ,
   ,a
,  ,
, ,, , ,    ,

Output:

a
aa
a,
,a
a a,
,a a
a,b
,a,b c,d
,a,b   c d   ef,g h
,
,a
,,
,,,,,,

The basic idea is:

• if we see a field of N spaces that doesn't contain any commas, followed by a character C which isn't a space or comma, then we just reproduce N spaces and character C.

• if we see a field of N spaces (possibly 0) containing one or more commas M, followed by a character C that isn't a space or comma, we reproduce the M commas, followed by C.

• lines in C streams are terminated by the newline character '\n' , which serves as C in the case when the comma-space field is the last item in the line.

A C program that doesn't manipulate any pointers cannot have a buffer overflow or memory leak. However, I haven't protected the counters against integer overflow. If you have a field of more than INT_MAX spaces and/or commas, the behavior is undefined. On modern systems, that's well over two billion, so there is a fair amount of justification for not caring about it.

The code also doesn't recognize other whitespace such as tabs.

Answer 3

char *trim(char *str, const char *chars)
{
    char *end = str + strlen(str) - 1;
    while(end > str)
    {
        if(strchr(chars, *str))
        {
            str++;
        }
        if(strchr(chars, *end))
        {
            *end-- = 0;
        }
    }
    return str;
}

and

comma = trim(comma, " ");