csv-pipe: read and write CSV in TypeScript, several times faster than papaparse

#node #javascript #webdev #typescript

A while back I wrote about the one-line fix for CSV formula injection. That fix ships in csv-pipe. Here is the rest of it.

csv-pipe is a typed, zero-dependency CSV library for TypeScript. It is one import for both directions, about 2 kB each way, and runs on Node, the browser, Deno, Bun, and edge.

Encode and parse are mirror images, so your rows round-trip:

import { stringify, parse } from 'csv-pipe';

type User = { name: string; age: number };

const users: User[] = [
  { name: 'Alex', age: 29 },
  { name: 'Carlos', age: 24 }
];

const csv = stringify(users);
// name,age
// Alex,29
// Carlos,24

parse<User>(csv, { dynamicTyping: true });
// [{ name: 'Alex', age: 29 }, { name: 'Carlos', age: 24 }]

Columns are checked against your type, so a typo is a compile error, not a broken export you find in production:

stringify(users, { columns: ['name', 'age'] }); // ok
stringify(users, { columns: ['naem'] });        // compile error: not a key of User

Parsing is the standout. On a 50,000-row dataset, in operations per second (higher is better):

Library	parse ops/sec
csv-pipe	201
papaparse	61
csv-parse	32
fast-csv	16

That is roughly 3x to 13x faster than papaparse, csv-parse, and fast-csv, with encoding about 1.5x ahead. These are indicative numbers from one machine, and the benchmark is in the repo, so you can reproduce them with npm run bench.

papaparse, csv-parse with csv-stringify, and fast-csv are all solid and worth knowing. Reach for csv-pipe when you want all of that in one small, typed package.

The real test is your own data, so paste a CSV into the playground:
https://martsinlabs.github.io/csv-pipe/playground

It is open source and still growing. If it is useful to you, a star helps others find it, and issues, ideas, and pull requests are all welcome.