Parsing error

[hanguanmiao]

problem1

int,string
1,

One less column is parsed, the last column should be null

problem2

int,string,int
1,"",123

There are 3 columns in total, but only 2 columns can be parsed, '",123' is treated as one column

[ljluestc]

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

// Maximum length of a CSV line
#define MAX_LINE_LENGTH 1024
#define MAX_COLUMNS 100  // Adjust based on expected column count

// Function to trim leading and trailing whitespace
char *trim_whitespace(char *str) {
    while (*str == ' ' || *str == '\t') str++; // Trim leading whitespace
    char *end = str + strlen(str) - 1;
    while (end > str && (*end == ' ' || *end == '\t' || *end == '\n' || *end == '\r')) end--; // Trim trailing whitespace
    *(end + 1) = '\0';
    return str;
}

// CSV Parsing Function
int parse_csv_line(const char *line, char *columns[], int max_columns) {
    int count = 0;
    const char *ptr = line;
    int in_quotes = 0;
    char buffer[MAX_LINE_LENGTH];
    int buf_pos = 0;

    while (*ptr) {
        if (*ptr == ',' && !in_quotes) {  // Column separator
            buffer[buf_pos] = '\0';
            columns[count++] = strdup(trim_whitespace(buffer)); // Store parsed field
            buf_pos = 0;
            if (count >= max_columns) break;
        } else if (*ptr == '"') {
            if (in_quotes && *(ptr + 1) == '"') {  // Handle escaped quotes
                buffer[buf_pos++] = '"';
                ptr++;  // Skip next quote
            } else {
                in_quotes = !in_quotes;  // Toggle quote state
            }
        } else {
            buffer[buf_pos++] = *ptr;
        }
        ptr++;
    }

    buffer[buf_pos] = '\0';
    columns[count++] = strdup(trim_whitespace(buffer));  // Store last column

    // Handle missing last column (empty column at the end)
    if (*(ptr - 1) == ',') {
        columns[count++] = strdup("");  // Add empty value for trailing comma
    }

    return count;
}

// Test the CSV parser
void test_csv_parser() {
    char *columns[MAX_COLUMNS];
    int col_count;

    const char *test_cases[] = {
        "1,",          // Problem 1: Last column should be NULL/empty
        "1,\"\",123",  // Problem 2: Correctly parse quoted empty strings
    };

    for (int i = 0; i < 2; i++) {
        printf("Parsing: %s\n", test_cases[i]);
        col_count = parse_csv_line(test_cases[i], columns, MAX_COLUMNS);

        for (int j = 0; j < col_count; j++) {
            printf("Column %d: '%s'\n", j + 1, columns[j]);
            free(columns[j]);  // Free allocated memory
        }
        printf("\n");
    }
}

int main() {
    test_csv_parser();
    return 0;
}