Cannot properly count elements in CSV after strtok in C

Question

I only sometimes see a proper count of the number of elements after I use strtok to delimit by commas and quotation marks (double quotes). Usually the printf is length 0, but occasionally 6 and 1 with no change to the code.

I have tried only using one delimiter (commas) and defining the token output of strtok in different ways, and rearranging the order of statements in the while loop where the other elements of the row are supposedly delimited. These are a few lines of the .csv file I am using to test my code (test.csv). It is the same format that NOAA provides precipitation data in as a .csv.

"STATION","NAME","DATE","PRCP","PRCP_ATTRIBUTES"
"US183459384","XYZ ABC 9.0 E, WA US","2019-01-06","0.65",",,N"
"US183459384","XYZ ABC 9.0 E, WA US","2019-01-12","0.46",",,N"
"US183459384","XYZ ABC 9.0 E, WA US","2019-01-13","0.09",",,N"
"US183459384","XYZ ABC 9.0 E, WA US","2019-01-14","0.01",",,N"
"US183459384","XYZ ABC 9.0 E, WA US","2019-01-15","0.60",",,N"
"US183459384","XYZ ABC 9.0 E, WA US","2019-01-16","1.93",",,N"

And my code attempt is below.

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

#define BUFFER_SIZE 1024

int get_row(FILE *file, int row_num, char delim[]) {
    int n_line = 0;
    int field = 0;
    char row[BUFFER_SIZE], *line[BUFFER_SIZE];

    while (fgets(row, BUFFER_SIZE, file)) {
        if (n_line == row_num) {
            printf("Length of line %d is %ld elements!\n", n_line, strlen(row));
            char* element = strtok(row, delim);
            while (element != NULL) {
                printf("%s\n", element);
                line[field++] = strdup(element);
                element = strtok(NULL, delim);
            }
            return 0;
        } else {
          n_line++;
    }
    printf("There is no row %d in the file you selected.\n", row_num);
    return 0;
}

int main(int argc, char **argv) {
    FILE *file;
    char delim[]  = ", \"";

    file = fopen(”test.csv”, "r");
    if (!file) {
        printf("Error: could not open %s\n", file_name);
        return -1;
    }
    printf("Reading file...\n");
    get_row(file, 0, delim);
    fclose(file);
    return 0;
}

I expect the result to show 5 but the result is 0 or 1 for all lines and occasionally 6.

https://ericlippert.com/2014/03/05/how-to-debug-small-programs/ — melpomene, Jul 01 '19 at 07:51
yes sorry that was originally an while(fgets(row, BUFFER_SIZE, file)) {. I mistyped when I transferred to SO — rc341, Jul 01 '19 at 07:56
You cannot use `strtok` to parse a CSV like this. Your input data contains commas and spaces, so `strtok` will split those fields at each comma and space. If your data does not contain quotes, you might get away with only using quotes as delimitors, and skipping every second field (which would be the commas between the actual fields). — rtoijala, Jul 01 '19 at 08:07
See e.g. [this tutorial](https://codingboost.com/parsing-csv-files-in-c) for hints — rtoijala, Jul 01 '19 at 08:32

score 1 · Accepted Answer · answered Jul 01 '19 at 08:54

This program should not compile as file_name is not defined. Also, inside the getrow function no of elements should be equal to field and not the length of the buffer. Also the comma delim will not work as there are commas in the field. The following code is working for the given lines in the test.csv file

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

#define BUFFER_SIZE 1024

int get_row(FILE *file, int row_num, char delim[]) {
    int n_line = 0;
    int field = 0;
    char row[BUFFER_SIZE], *line[BUFFER_SIZE];

    while (fgets(row, BUFFER_SIZE, file)) {
        if (n_line == row_num) {
            char* element = strtok(row, delim);
            while (element != NULL) {
                if(strcmp(",", element) != 0 &&  strcmp("\n", element) != 0)
                {
                  printf("%s\n", element);
                  line[field++] = strdup(element);
                }
                element = strtok(NULL, delim);
            }
            printf("Length of line %d is %d elements!\n", n_line, field);
            return 0;
        } else {
          n_line++;
    }
  }
    printf("There is no row %d in the file you selected.\n", row_num);
    return 0;
}

int main(int argc, char **argv) {
    FILE *file;
    char delim[]  = "\"";
    char file_name[] = "test.csv";

    file = fopen(file_name, "r");
    if (!file) {
        printf("Error: could not open %s\n", file_name);
        return -1;
    }
    printf("Reading file...\n");
    get_row(file, 0, delim);
    fclose(file);
    return 0;
}

Alex DiS · Answer 2 · 2019-07-01T08:44:26.413

0

Many syntax errors in program. This should work:

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

#define BUFFER_SIZE 1024

int get_row(FILE *file, int row_num, char delim[]) {
    char row[BUFFER_SIZE];

    for (int i = 0; fgets(row, BUFFER_SIZE, file); i++) {
        if (i == row_num) {
            printf("Length of line %d is %ld elements!\n", i, strlen(row));

            char* element = strtok(row, delim);
            while (element) {
                printf("%s\n", element);
                element = strtok(NULL, delim);
            }

            break;
        }
    }

    printf("There is no row %d in the file you selected.\n", row_num);

    return 0;

}

int main() {
    FILE *file;
    char delim[]  = ", \"";

    file = fopen("test.csv", "r");
    if (!file) {
        puts("Error: could not open");
        return -1;
    }

    printf("Reading file...\n");
    get_row(file, 0, delim);
    fclose(file);

    return 0;
}

edited Jul 01 '19 at 08:44

answered Jul 01 '19 at 08:13

Alex DiS

49
7

This does not deal with commas inside fields. – rtoijala Jul 01 '19 at 08:20
@rtoijala fix it – Alex DiS Jul 01 '19 at 08:34
The edit does not fix the problem. Now you have incorrect output for any line containing commas, spaces, or double quotes inside any field. Change the `0` to `1` in the call to `get_row` to see what I mean. – rtoijala Jul 01 '19 at 08:38

Cannot properly count elements in CSV after strtok in C

2 Answers2