@kevin-coelho/json-arraystream v1.0.1

ArrayStream

ArrayStream is a node.js utility to stream large quantities of json data to file, formatted as a large JSON array. These can later be read using utilities like https://www.npmjs.com/package/stream-json.

Features

• Extends node Writeable
• Will automatically convert JS objects into valid JSON
• Validation on each JSON chunk
• Automatically creates output folder and uses a basic increment file naming strategy
• Wait for stream end using bluebird Promise
• High level metadata / statistic such as: total count of items written, list of filenames written to
• Option to append to or overwrite existing files.
• Option to avoid existing files and start from nearest filename increment (if file_0001.json and file_0002.json exist, write to file_0003.json)
• Set maximum number of objects to be written for each file and ArrayStream will auto-increment filename
• Option to set stream to "lazy" mode and not open any file handles until needed

Usage

const ArrayStream = require('@kevin-coelho/json-arraystream');

// see lib/types.js for config definition
const config = {
	file: {
		folder: './output',
		keyPattern: 'testFile_$$',
		fileType: 'json',
		numDigits: 4, // will create files like output/testFile_0001.json
	},
	maxItems: 10,
	itemValidationSchema: joi
		.object()
		.keys({
			someNumber: joi.number().required(),
		})
		.required(),
	lazy: true,
};

const out = new ArrayStream(config);
out.on('error', () => console.error('an error occurred'));

// write like a normal stream
out.write({
	someNumber: 5
});

// stream produces joi validation error when chunk fails to pass itemValidationSchema
out.write({
	foo: 'bar',
});

// out will create a new file testFile_0001.json, testFile_0002.json, ... for every 10 chunks written
const hugeArray = [....];
hugeArray.forEach(chunk => out.write(chunk));

await out.closeArrayStream();

Installation & Test

• npm install @kevin-coelho/json-arraystream
• npm run test

Roadmap

• Support limiting files by size rather than item limit
• Support streaming output to AWS S3 instead of local file