node-jq is a Node.js wrapper for jq - a lightweight and flexible command-line JSON processor

$ npm install node-jq --save
# or
$ yarn add node-jq

You can use jq directly after installing it:

npx node-jq '.foo' package.json

By default, node-jq downloads jq during the installation process with a post-install script. Depending on your SO downloads from [https://github.com/jqlang/jq/releases] into ./node_modules/node-jq/bin/jq to avoid colisions with any global installation. Check #161 #167 #171 for more information. You can safely rely on this location for your installed jq, we won't change this path without a major version upgrade.

If you want to skip the installation step of jq, you can set NODE_JQ_SKIP_INSTALL_BINARY to true or ignore the post-install script from the installation npm install node-jq --ignore-scripts.

export NODE_JQ_SKIP_INSTALL_BINARY=true
npm install node-jq

npm install node-jq --ignore-scripts

Usually in your CLI using jq:

jq ".abilities[].moves" bulbasaur.json

and you get

{
  "name": "heartgold-soulsilver",
  "power": "10"
}
{
  "name": "platinum",
  "power": "50"
}
{
  "name": "diamond-pearl",
  "power": "99"
}

With node-jq you could run it programmatically and interact with the output as a JavaScript Object:

NOTE: Take care of the filter that you are using with jq, mapping an array or any other iterative output isn't a valid JavaScript Object, that might fail at parse-time.

const jq = require('node-jq')

const filter = '.abilities[].moves'
const jsonPath = '/path/to/bulbasaur.json'
const options = {}

jq.run(filter, jsonPath, options)
  .then((output) => {
    console.log(output)
    /*
      {
        "name": "heartgold-soulsilver",
        "power": "10"
      },
      {
        "name": "platinum",
        "power": "50"
      },
      {
        "name": "diamond-pearl",
        "power": "99"
      }
    */
  })
  .catch((err) => {
    console.error(err)
    // Something went wrong...
  })

By default, the jq binary installed with the package is used. If you have special needs or want to use another binary in a different path you can set the environment variable JQ_PATH to override the binary path.

Run the jq query against a JSON file.

jq.run('.', '/path/to/file.json').then(console.log)
// { "foo": "bar" }

Run jq query against multiple JSON files.

jq.run('.', ['/path/to/file.json','path/to/other_file.json']).then(console.log)
// { "foo": "bar" }
// { "otherFoo": "andBar" }

Run the jq query against an Object.

jq.run('.', { foo: 'bar' }, { input: 'json' }).then(console.log)
// { "foo": "bar" }

Run the jq query against a String.

jq.run('.', '{ foo: "bar" }', { input: 'string' }).then(console.log)
// { "foo": "bar" }

Return the output as a String.

jq.run('.', '/path/to/file.json', { output: 'string' }).then(console.log)
// {
//   "foo": "bar"
// }

Return the output as an Object.

jq.run('.', '/path/to/file.json', { output: 'json' }).then(console.log)
// { foo: 'bar' }

Return the output as a String.

jq.run('.', '/path/to/file.json', { output: 'compact' }).then(console.log)
// {"foo":"bar"}
jq.run('.', '/path/to/file.json', { output: 'string' }).then(console.log)
// {"foo":"bar"}

Read input stream into array.

jq.run('.', ['/path/to/file.json','/path/to/other_file.json'], { output: 'json', slurp: true }).then(console.log)
// [
//   {
//     "foo": "bar"
//   },
//   {
//     "otherFoo": "andBar"
//   }
// ]

Sorts object keys alphabetically.

jq.run('.', ['/path/to/file.json'], { output: 'json', sort: true }).then(console.log)
// {
//   "a": 2,
//   "b": 1
// },

Adds the --argjson myfruit "{ 'hello': 'orange' }" --arg myfruit2 orange arguments to the internal jq command

jq.run('{"fruit":$myfruit,"fruit2":$myfruit2}', ['/path/to/file.json'], { output: 'json', sort: true, args: { myfruit: { hello: 'orange' }, myfruit2: "banana" } }).then(console.log)
// {
//   fruit: { 
//      hello: "orange" 
//   },
//   fruit2: "banana"
// }

jq.run('.', ['file.json'], { output: 'json', sort: true }, '/path/to').then(console.log)
// {
//   "a": 2,
//   "b": 1
// },

By default jq process will run 'attached' to the main process. That means that any interrupt signal main process receives will be propagated to jq process. For example, if main process receives SIGTERM, jq will also receive it and exit immediately.

However, in some cases you might not want jq to exit immediately and let it exit normally. For example, if you want to implement a graceful shutdown - main process receives SIGTERM, it finishes processing current json file and exits after processing is completed.

To achieve that run jq detached and NodeJS will not propagate SIGTERM to jq process allowing it to run until it completes.

jq.run('.', ['file.json'], { output: 'json', sort: true }, undefined, true).then(console.log)
// {
//   "a": 2,
//   "b": 1
// },

• atom-jq: an Atom package for manipulating JSON
• json-splora: an Electron implementation for manipulating JSON
• Check more

Why would you want to manipulate JavaScript Objects with jq inside a nodejs app, when there are tools like ramda or lodash?

The idea was to port jq in node to be able to run it as-is. node-jq doesn't try to replace Array/Object filters, maps, transformations, and so on.

Our primary goal was to make jq syntax available inside an Atom extension: atom-jq.

Other than that, jq is an interesting CLI tool to quickly parse and manipulate the response of an API, such as:

curl 'https://jsonplaceholder.typicode.com/comments' | jq '.[].postId'

There are also people dealing with complex use-cases, and some of them want to port their bash scripts to node:

• ilya-sher.org/2016/05/11/most-jq-you-will-ever-need
• cloudadvantage.com.au/new-aws-command-line-tool-and-jq

Seems hard to learn, but it really isn't.

jq is like sed for JSON. Slice, filter, map and transform structured data in a simple and powerful way.

Take a look at this great introduction or a jq lesson.

You can check out the official manual and fiddle around in the online playground jqplay.org.

License