Got it! What you're asking for is to track the length of each value (including leading/trailing whitespace, quotes, etc.) between delimiters, while also handling delimiters inside quoted values properly.

Here's how we can achieve that:

Parse the CSV with PapaParse (which correctly handles quoting).
Reconstruct each value's length based on the raw text between delimiters.

This approach will involve:

Parsing the CSV using PapaParse.
Iterating through the parsed rows and recording the length of each value as it appears in the raw text, including quotes, whitespace, etc.

Here's a solution that does this:

let csvContent = `name,age,city
"John, Doe", 30 ,"New York"
Jane, 25, "Los Angeles"`;

// Step 1: Parse the CSV to extract the values, and infer the delimiter
let config = {
    delimiter: "", // Allows PapaParse to infer the delimiter
    skipEmptyLines: false,
    keepEmptyRows: true, // If you want to keep empty rows
};

// Parse the CSV data
let result = Papa.parse(csvContent, config);
let delimiter = result.meta.delimiter;

// Step 2: Reconstruct raw text and capture value lengths
let valueLengths = [];
let rows = csvContent.split(/\r?\n/); // Split raw content by newlines

rows.forEach((row) => {
    let inQuotes = false;
    let currentValue = '';
    let valueStart = 0;
    let rowLengths = [];

    for (let i = 0; i < row.length; i++) {
        let char = row[i];

        // Toggle inQuotes when encountering double quotes
        if (char === '"') {
            inQuotes = !inQuotes;
        }

        // If we're not in quotes and hit a delimiter, calculate the value length
        if (char === delimiter && !inQuotes) {
            rowLengths.push(i - valueStart); // Length from start to current delimiter
            valueStart = i + 1; // Start of next value
        }
    }

    // After the last delimiter, push the final value's length
    rowLengths.push(row.length - valueStart);

    // Collect the row's value lengths
    valueLengths.push(rowLengths);
});

console.log("Delimiter:", delimiter);
console.log("Value lengths per row:", valueLengths);

How it works:

Step 1: Use PapaParse to correctly parse the CSV and infer the delimiter.
Step 2: Split the original CSV content by lines, and manually scan through each line.
- Track quoted sections so delimiters inside quotes are ignored.
- Each time a delimiter is encountered outside of quotes, the length of the value (including whitespace and quotes) is recorded.
- After processing each line, record the lengths of the values.

Example Output:

For the input:

name,age,city
"John, Doe", 30 ,"New York"
Jane, 25, "Los Angeles"

You would get something like:

Delimiter: ,
Value lengths per row: 
[
  [4, 3, 4], // For "name,age,city"
  [11, 5, 10], // For '"John, Doe", 30 ,"New York"'
  [4, 3, 13]  // For 'Jane, 25, "Los Angeles"'
]

This gives you the exact length of the raw values between the delimiters, including any quotes and whitespace.

Let me know if this works for your needs or if you'd like further refinements!

Use a CSV parser #1

Description

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions