batch-request-js

Batch promise based requests to overcome network limitations or API restrictions

Install

• yarn add batch-request-js

Tests

• yarn test-unit to run unit tests
• yarn test-e2e to run e2e tests
• yarn test-ci to run both unit and e2e tests

Usage

Suppose we'd like to fetch thousands of customers from an API. To avoid network limitations or rate limiting issues, we can batch the requests:

// node.js
const batchRequest = require('batch-request-js')


async function getCustomers () {
  // define array of input data e.g. customerIds
  const customerIds = ['100', '101', '102', ... ]
  // define the async request to perform against each data input
  const request = (customerId) => fetch(`${API_ENDPOINT}/${customerId}`).then(response => response.json())

  // fetch customers in batches of 100, delaying 200ms inbetween each batch request
  const { error, data } = await batchRequest(customerIds, request, { batchSize: 100, delay: 200 })

  // Data from successful requests
  console.log(data) // [{ customerId: '100', ... }, ...]

  // Failed requests
  console.log(error) // [{ record: "101", error: [Error: Customer not found] }, ...]
}

Example

// node.js
const batchRequest = require('batch-request-js')

async function getData () {
  // setup a 100 test records
  const records = Array(100).fill(0).map((d, i) => ({ item: i }))
  // all requests will succeed and be timestamped
  const request = record => Promise.resolve({ ...record, timestamp: Date.now() })
  // batch requests 20 at a time, delaying half a second after each batch request
  const result = await batchRequest(records, request, { batchSize: 20, delay: 500 })
  console.log(result)
//   { 
//     error: [],
//     data: [
//        { record: 0, timestamp: 1533552890663 },
//        { record: 1, timestamp: 1533552890663 },
//        { record: 2, timestamp: 1533552890663 },
//        { record: 3, timestamp: 1533552890663 },
//         ...
//     ]
}

getData()

Handling failed batch requests

Rerun batch request with a filtered set of inputRecords to just those that failed on the previous attempt.

Future

Retry logic may be implemented to handle automatically rerunning batch-request for failing requests.

License

MIT