Javascript CSV Strings

Parse and Stringify for CSV strings.

• API similar to the JSON parser (CSV.parse and CSV.stringify).
• Can also work row by row.
• Can also be used to parse strings from readable streams (e.g. file streams).
• Tolerant with the weird data
• Written in TypeScript

import * as CSV from 'csv-string';

// with String
const arr = CSV.parse('a,b,c\na,b,c');
const str = CSV.stringify(arr);

// with Stream
const stream = CSV.createStream();
stream.on('data', rows => {
  process.stdout.write(CSV.stringify(rows, ','));
});
process.stdin.pipe(stream);

Contributors

• Kael Shipman
• Mehul Mohan
• Hossam Magdy
• Rich
• Rick Huizinga
• Nicolas Thouvenin
• Stéphane Gully
• J. Baumbach
• Sam Hauglustaine
• Rick Huizinga
• doleksy1
• François Parmentier

Installation

using npm:

npm install csv-string

or yarn

yarn add csv-string

API Documentation

parse(input: String, [options: Object]): Object

parse(input: string, [separator: string], [quote: string]): Object

Converts a CSV string input to array output.

Options :

• comma String to indicate the CSV separator. (optional, default ,)
• quote String to indicate the CSV quote if need. (optional, default ")
• output String choose 'objects' or 'tuples' to change output for Array or Object. (optional, default tuples)

Example 1 :

const CSV = require('csv-string');
const parsedCsv = CSV.parse('a;b;c\nd;e;f', ';');
console.log(parsedCsv);

Output:

[
  ["a", "b", "c"],
  ["d", "e", "f"]
]

Example 2 :

const CSV = require('csv-string');
const parsedCsv = CSV.parse('a,b,c\n1,2,3\n4,5,6', { output: 'objects' });
console.log(parsedCsv);

Output:

[
  { a: '1', b: '2', c: '3' },
  { a: '4', b: '5', c: '6' }
]

If separator parameter is not provided, it is automatically detected.

stringify(input: Object, [separator: string]): string

Converts object input to a CSV string.

import * as CSV from 'csv-string';

console.log(CSV.stringify(['a', 'b', 'c']));
console.log(
  CSV.stringify([
    ['c', 'd', 'e'],
    ['c', 'd', 'e']
  ])
);
console.log(CSV.stringify({ a: 'e', b: 'f', c: 'g' }));

Output:

a,b,c

c,d,e
c,d,e

e,f,g

detect(input: string): string

Detects the best separator.

import * as CSV from 'csv-string';

console.log(CSV.detect('a,b,c'));
console.log(CSV.detect('a;b;c'));
console.log(CSV.detect('a|b|c'));
console.log(CSV.detect('a\tb\tc'));

Output:

,
;
|
\t

forEach(input: string, sep: string, quo: string, callback: function)

forEach(input: string, sep: string, callback: function)

forEach(input: string, callback: function)

callback(row: array, index: number): void

Calls callback for each CSV row/line. The Array passed to callback contains the fields of the current row.

import * as CSV from 'csv-string';

const data = 'a,b,c\nd,e,f';

CSV.forEach(data, ',', function (row, index) {
  console.log('#' + index + ' : ', row);
});

Output:

#0 :  [ 'a', 'b', 'c' ]
#1 :  [ 'd', 'e', 'f' ]

read(input: string, sep: string, quo: string, callback: function): number

read(input: string, sep: string, callback: function): number

read(input: string, callback: function): number

callback(row: array): void

Calls callback when a CSV row is read. The Array passed to callback contains the fields of the row. Returns the first offset after the row.

import * as CSV from 'csv-string';

const data = 'a,b,c\nd,e,f';

const index = CSV.read(data, ',', row => {
  console.log(row);
});

console.log(data.slice(index));

Output:

[ 'a', 'b', 'c' ]
d,e,f

readAll(input: string, sep: string, quo: string, callback: function): number

readAll(input: string, sep: string, callback: function): number

readAll(input: string, callback: function): number

callback(rows: array): void

Calls callback when all CSV rows are read. The Array passed to callback contains the rows of the file. Returns the offset of the end of parsing (generally it's the end of the input string).

import * as CSV from 'csv-string';

const data = 'a,b,c\nd,e,f';

const index = CSV.readAll(data, row => {
  console.log(row);
});

console.log('-' + data.slice(index) + '-');

Output:

[ [ 'a', 'b', 'c' ], [ 'd', 'e', 'f' ] ]
--

readChunk(input: string, sep: string, quo: string, callback: function): number

readChunk(input: string, sep: string, callback: function): number

readChunk(input: string, callback: function): number

callback(rows: array): void

Calls callback when all CSV rows are read. The last row could be ignored, because the remainder could be in another chunk. The Array passed to callback contains the rows of the file. Returns the offset of the end of parsing. If the last row is ignored, the offset will point to the beginnning of the row.

import * as CSV from 'csv-string';

const data = 'a,b,c\nd,e';

const index = CSV.readChunk(data, row => {
  console.log(row);
});

console.log('-' + data.slice(index) + '-');

Output:

[ [ 'a', 'b', 'c' ] ]
--

createStream(options: Object): WritableStream

createStream(): WritableStream

Create a writable stream for CSV chunk. Options are :

• separator : To indicate the CSV separator. By default is auto (see the detect function)
• quote** : To indicate the CSVquote.

Example : Read CSV file from the standard input.

const stream = CSV.createStream();

stream.on('data', row => {
  console.log(row);
});

process.stdin.resume();
process.stdin.setEncoding('utf8');
process.stdin.pipe(stream);

Contribution

• clone
• yarn install
• ... do the changes, write tests
• yarn test (ensure all tests pass)
• yarn bench (to check the performance impact)

Related projects

• https://npmjs.org/browse/keyword/csv
• http://www.uselesscode.org/js/csv/
• https://github.com/archan937/csonv.js

Benchmark

There is a quite basic benchmark to compare this project to other related ones, using file streams as input. See ./bench for source code.

the test

yarn bench

the result

for a test file with 949,044 rows

Package	Time	Output/Input similarity
a-csv	6.01s	~99%
csv-stream	6.64s	~73%
csv-streamer	7.03s	~79%
csv-string	6.53s	100%
fast-csv	12.33s	99.99%
nodecsv	7.10s	100%

License

MIT/X11