Convert array of objects into a CSV file

Converts an array of JavaScript objects into the CSV format. You can save the CSV to file or return it as a string.

The keys in the first object of the array will be used as column names.

Any special characters in the values (such as commas) will be properly escaped. Spreadsheet formula escaping is enabled by default for both values and column names.

Upgrading from 1.x

Version 3 requires Node.js 18 or newer. Existing toString() and toDisk() calls still work, but strings that could be interpreted as spreadsheet formulas now get a leading apostrophe. For example, =1+1 becomes '=1+1, and the string -42 becomes '-42. JavaScript numbers such as -42 remain numbers in the CSV output. Null rows, array rows and sparse input arrays are rejected with a clear error.

The version number skips the previously published experimental 2.0.0 on the next npm tag. This release continues the API from stable 1.3.6.

Usage

const ObjectsToCsv = require('objects-to-csv');

// Sample data - two columns, three rows:
const data = [
  {code: 'CA', name: 'California'},
  {code: 'TX', name: 'Texas'},
  {code: 'NY', name: 'New York'},
];

// If you use "await", code must be inside an asynchronous function:
(async () => {
  const csv = new ObjectsToCsv(data);

  // Save to file:
  await csv.toDisk('./test.csv');

  // Return the CSV file as string:
  console.log(await csv.toString());
})();

Methods

There are two methods, toDisk(filename) and toString().

new ObjectsToCsv(data, options)

data must be an array of non-null objects. The optional options object accepts:

• escapeFormulas — defaults to true. Prefixes potentially dangerous text with an apostrophe before CSV quoting. Applies to column names and data in both output methods, including appended rows. Does not mutate the input objects.

Protection covers =, +, -, @, their full-width Unicode variants, and leading tabs, carriage returns and line feeds. Formula prefixes behind leading whitespace or ASCII control characters are also escaped. Existing leading apostrophes are preserved.

Escaping changes the underlying text. To preserve raw values for trusted data or machine-only processing, explicitly disable it:

const csv = new ObjectsToCsv(trustedData, { escapeFormulas: false });

With escaping disabled, untrusted values can become active spreadsheet formulas. CSV quoting alone does not prevent this. Spreadsheet import settings differ, and editing, saving and reopening CSV files can remove protective escaping. For a workflow that must preserve cell types through editing, use a spreadsheet format with explicitly typed text cells. See OWASP's CSV injection guidance.

async toDisk(filename, options)

Converts the data and saves the CSV file to disk. The filename must include the path as well.

The options is an optional parameter which is an object that contains the settings. Supported options:

• append - whether to append to the file. Default is false (overwrite the file). Set to true to append. Column names will be added only once at the beginning of the file. If the file does not exist, it will be created.
• bom - whether to add the Unicode Byte Order Mark at the beginning of the file. Default is false; set to true to be able to view Unicode in Excel properly. Otherwise Excel will display Unicode incorrectly. The BOM is included when creating or overwriting a file; appending to a non-empty file does not add another BOM.
• allColumns - whether to check all array items for keys to convert to columns rather than only the first. This will sort the columns alphabetically. Default is false; set to true to check all items for potential column names.

const ObjectsToCsv = require('objects-to-csv');
const sampleData = [{ id: 1, text: 'this is a test' }];

// Run asynchronously, without awaiting:
new ObjectsToCsv(sampleData).toDisk('./test.csv');

// Alternatively, you can append to the existing file:
new ObjectsToCsv(sampleData).toDisk('./test.csv', { append: true });

// `allColumns: true` collects column names from all objects in the array,
// instead of only using the first one. In this case the CSV file will
// contain three columns:
const mixedData = [
  { id: 1, name: 'California' },
  { id: 2, description: 'A long description.' },
];
new ObjectsToCsv(mixedData).toDisk('./test.csv', { allColumns: true });

Resolves to the text written to the file. Await each write before appending again to the same file. When appending, keep the same column names and order as the existing CSV; the method does not reconcile schemas.

async toString(header = true, allColumns = false)

Returns the CSV file as a string.

Two optional parameters are available:

• header controls whether the column names will be returned as the first row of the file. Default is true. Set it to false to get only the data rows, without the column names.
• allColumns controls whether to check every item for potential keys to process, rather than only the first item; this will sort the columns alphabetically by key name. Default is false. Set it to true to process keys that may not be present in the first object of the array.

const ObjectsToCsv = require('objects-to-csv');
const sampleData = [{ id: 1, text: 'this is a test' }];

async function printCsv(data) {
  console.log(
    await new ObjectsToCsv(data).toString()
  );
}

printCsv(sampleData);

Requirements

Use Node.js version 18 or above. A currently maintained Node.js release is recommended.

Development

npm ci
npm test
npm audit
npm pack --dry-run

Tests use Node's built-in test runner. CSV formatting is handled by csv-stringify, the only runtime dependency. Publishing runs the tests again through prepublishOnly.