C: Combining `sscanf` specifiers to read upto a number of characters

Question

I need to read lines of strings in a text file that represent movie showings and format them. I need to use sscanf to scan the string saved by fgets. My problem is how do I make sscanf only read upto x amount of characters while also using [^] specifier. Movie title lengths have a max length of 44. I know C has %0.*s but I need to use it in combination with [^]. I tried doing %0.44[^,] but to no avail. My code is below. I have commented out what I though would be the solution.

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

int main(void) {
    const int MAX_TITLE_CHARS = 44;  // Maximum length of movie titles
    const int LINE_LIMIT = 100;   // Maximum length of each line in the text file
    char line[LINE_LIMIT];
    char inputFileName[25];


    FILE *file;
    file = fopen("D:\\movies.csv", "r");

    char currentLine[LINE_LIMIT];
    char movieTitle[MAX_TITLE_CHARS];
    char movieTime[5];
    char movieRating[5];

    fgets(currentLine, LINE_LIMIT, file);
    while(!feof(file)){
        
    //  sscanf(currentLine, "%[^,],%0.44[^,],%[^,]", movieTime, movieTitle, movieRating);
        sscanf(currentLine, "%[^,],%[^,],%[^,]", movieTime, movieTitle, movieRating);

        printf("%-44s |\n", movieTitle);
        fgets(currentLine, LINE_LIMIT, file);
    }


    return 0;
}

This prints out the following

Wonders of the World                         |
Wonders of the World                         |
Journey to Space                             |
Buffalo Bill And The Indians or Sitting Bull's History Lesson |
Buffalo Bill And The Indians or Sitting Bull's History Lesson |
Buffalo Bill And The Indians or Sitting Bull's History Lesson |
Adventure of Lewis and Clark                 |
Adventure of Lewis and Clark                 |
Halloween                                    |

I need to it be

Wonders of the World                         |
Wonders of the World                         |
Journey to Space                             |
Buffalo Bill And The Indians or Sitting Bull |
Buffalo Bill And The Indians or Sitting Bull |
Buffalo Bill And The Indians or Sitting Bull |
Adventure of Lewis and Clark                 |
Adventure of Lewis and Clark                 |
Halloween                                    |

See [Why is “while ( !feof (file) )” always wrong?](https://stackoverflow.com/q/5431941/2410359). This code fails on input error. — chux - Reinstate Monica, Mar 11 '22 at 13:32
Petetunze, what should happen if the line read has a time, title, rating exceeding 5, 44, 5 characters? — chux - Reinstate Monica, Mar 11 '22 at 13:35

score 1 · Accepted Answer · edited Mar 11 '22 at 12:41

1

sscanf(currentLine, "%[^,],%43[^,],%[^,]", movieTime, movieTitle, movieRating);

Where 43 is the amount of chars to read and one left for a null terminator. I assume the compiler was unhappy about the 0. complaining about zero width and unknown conversion type.

edited Mar 11 '22 at 12:41

vvvvv

25,404
19
49
81

answered Mar 11 '22 at 12:05

Gerhard

6,850
8
51
81

1

Since the array is 44 characters long, you need 43 in the format string to allow for the thanks byte. The “off by one” is an unfortunate but sacrosanct historical artefact. – Jonathan Leffler Mar 11 '22 at 12:08
@Jonathan_Leffler: Correct. I did not double check the buffer size in the question. – Gerhard Mar 11 '22 at 12:10
I wrote “null byte” but my spell mangler angles null to thanks for some reason and I didn't spot it's subversive tactic until too late. – Jonathan Leffler Mar 11 '22 at 12:18
`%[^,]` risks buffer overflow. Use a width limits there too. – chux - Reinstate Monica Mar 11 '22 at 13:28
Since OP's goal included "Movie title lengths have a max length of 44.", makes more sense to use width of 44 and increate the buffer size than restrict the width to 43. – chux - Reinstate Monica Mar 11 '22 at 13:36

Eric Postpischil · Answer 2 · 2022-03-11T12:18:15.967

I know C has %0.*s…

No, it does not. For sscanf, the way to limit the number of characters accepted is to give the width as a decimal numeral after the %, without 0 or ., as in "%3s". It must be an actual numeral; it cannot be * as in printf, which specifies that the number is passed as an argument.

To use this with the [ conversion specifier, you would use "%44[^,]".

If the maximum move title length is 44 characters, then the array that holds them as a string should be declared as char movieTitle[MAX_TITLE_CHARS+1]; to allow for a terminating null character.

If you want to parameterize the sscanf string based on the symbol MAX_TITLE_CHARS, so that it adapts if the value is changed, you can do this by defining MAX_TITLE_CHARS as a macro instead of a const int object:

#define MAX_TITLE_CHARS 44

and defining macros to convert a parameter into a string:

//  Two macros are needed due to the order of operations in macro replacement.
#define ExpandAndStringize(x)   #x
#define Stringize(x)            ExpandAndStringize(x)

and using them:

sscanf(currentline,"%[^,],%" Stringize(MAX_TITLE_CHARS) "[^,],%[^,]", movieTime, movieTitle, movieRating);

To avoid overflow, consider using width limits on the other 2 `"%[^,]` also. — chux - Reinstate Monica, Mar 11 '22 at 13:30

C: Combining `sscanf` specifiers to read upto a number of characters

2 Answers2