CSV Format: A Developer's Complete Guide

What is CSV?

CSV (Comma-Separated Values) is a plain-text format for tabular data. Each line represents a row, and fields within a row are separated by a delimiter — usually a comma. Despite its name, semicolons, tabs, and pipes are all common alternatives depending on locale and tooling.

CSV has no official standard, but RFC 4180 is the closest thing to a spec and is widely referenced. Most modern parsers follow it with varying degrees of strictness.

The RFC 4180 Rules

• Each record is on its own line, terminated by CRLF (\r\n)
• The last record may or may not have a trailing line break
• An optional header row appears as the first record
• Fields may be enclosed in double quotes
• A field containing commas, double quotes, or newlines must be enclosed in double quotes
• A double quote inside a quoted field is escaped by doubling it: ""

Valid CSV Example

id,name,notes
1,Alice,"Works in ""New York"""
2,Bob,"Line 1
Line 2"
3,Charlie,No quotes needed

Common Delimiters

The comma is the default, but real-world CSV often uses other delimiters:

• Semicolon (;) — Default in European Excel exports where commas are decimal separators
• Tab (\t) — Called TSV (Tab-Separated Values); preferred for data that often contains commas
• Pipe (|) — Used in logs, database dumps, and legacy systems

Always detect the delimiter programmatically when parsing unknown files — count occurrences of each candidate in the first line to make an educated guess.

CSV vs JSON

Both formats are ubiquitous, but they serve different purposes:

When to use CSV

• Exporting data to Excel, Google Sheets, or other spreadsheet tools
• Database imports and exports (COPY in PostgreSQL, LOAD DATA in MySQL)
• Large flat datasets where file size matters (CSV has no key repetition overhead)
• Sharing data with non-developers who will open it in a spreadsheet

When to use JSON

• Nested or hierarchical data (CSV cannot represent trees natively)
• REST APIs and web applications
• When type information matters (JSON preserves numbers, booleans, null)
• Configuration files and structured documents

Converting CSV to JSON

The basic algorithm for converting CSV with a header row to a JSON array of objects:

// 1. Parse CSV into rows (array of arrays)
const rows = parseCSV(csvText);

// 2. First row becomes the keys
const headers = rows[0];

// 3. Remaining rows become objects
const data = rows.slice(1).map(row =>
  Object.fromEntries(headers.map((h, i) => [h, row[i] ?? '']))
);

The tricky part is the parser itself — a naive split(',') breaks on quoted fields. A robust parser must track whether it is inside a quoted field character by character.

Common Pitfalls

• BOM characters — Excel adds a UTF-8 BOM (\uFEFF) that corrupts the first header name. Strip it before parsing.
• Windows line endings — \r\n vs \n. Normalize before splitting by lines.
• Trailing newline — A final empty line creates a phantom empty row. Drop it after parsing.
• Type coercion — CSV is always strings. "123" is not the number 123. Parse types explicitly after conversion.
• Inconsistent column counts — Some rows may have fewer fields than the header. Treat missing fields as empty strings.

Performance Tips for Large Files

• Stream the file rather than reading it all into memory
• In the browser, use FileReader.readAsText() for upload and parse in chunks
• For server-side Node.js, use a streaming CSV library (e.g., csv-parse in stream mode)
• Paginate the rendered table — rendering 100k rows at once in the DOM will freeze the browser