@arraypress/csv

CSV generation for edge runtimes. String output for small datasets, streaming for large ones, and a ready-made Response helper for Cloudflare Workers and Hono routes.

Zero dependencies. Uses Web Streams API — works in Cloudflare Workers, Node.js 18+, Deno, Bun, and browsers.

Installation

npm install @arraypress/csv

Usage

Quick Export (small datasets)

import { csvResponse } from '@arraypress/csv';

app.get('/api/export/customers', async (c) => {
  const data = await db.prepare('SELECT * FROM customers').all();

  return csvResponse({
    filename: 'customers.csv',
    columns: ['ID', 'Email', 'Name', 'Country', 'Orders'],
    rows: data.results.map(c => [c.id, c.email, c.name, c.country, c.order_count]),
  });
});

Streaming Export (large datasets)

import { csvResponse } from '@arraypress/csv';

app.get('/api/export/orders', async (c) => {
  return csvResponse({
    filename: 'orders.csv',
    columns: ['Order #', 'Email', 'Amount', 'Currency', 'Status', 'Date'],
    stream: async function* () {
      let offset = 0;
      while (true) {
        const batch = await db.prepare('SELECT * FROM orders LIMIT 500 OFFSET ?')
          .bind(offset).all();
        if (!batch.results.length) break;
        for (const o of batch.results) {
          yield [o.order_number, o.email, o.amount, o.currency, o.status, o.created_at];
        }
        offset += 500;
      }
    },
  });
});

Build CSV String

import { createCSV } from '@arraypress/csv';

const csv = createCSV({
  columns: ['Name', 'Email', 'Amount'],
  rows: [
    ['Alice', '[email protected]', 1999],
    ['Bob', '[email protected]', 2500],
  ],
});
// "Name,Email,Amount\nAlice,[email protected],1999\nBob,[email protected],2500\n"

API

csvResponse({ filename, columns, rows?, stream? })

Create a complete Response with CSV data, Content-Type, and Content-Disposition headers. Ready to return from any Worker/Hono route.

• filename — download filename (e.g. 'orders.csv')
• columns — header row
• rows — array of row arrays (for small datasets)
• stream — async generator yielding row arrays (for large datasets)
• includeHeader — include header row (default true)

If both rows and stream are provided, stream takes precedence.

createCSV({ columns, rows })

Returns a CSV string. Best for small datasets or when you need the string for something other than a Response (email attachment, file write, etc.).

createCSVStream({ columns, stream })

Returns a ReadableStream. Best for large datasets where you need the stream directly (piping to another stream, etc.).

Escaping

All values are automatically escaped per RFC 4180:

• Values containing commas are wrapped in quotes: "Smith, John"
• Values containing quotes have quotes doubled: "Say ""hello"""
• Values containing newlines are wrapped in quotes
• Null and undefined become empty strings
• Numbers and booleans are converted to strings

License

MIT