Yankee Doodle

Yankee Doodle, abbreviated as Yank, is a Lodash.pick inspired JavaScript package to pick object values from plain JavaScript objects. There’s zero dependencies to worry about, so naturally this package is really small at only around 9K minified. It’s safe to use in both browser and Node environments.

IE11 support is currently lacking due to some regular expressions that aren’t available in IE11 - stop using IE11!

Installing

$ npm i -S yankee-doodle

Getting Started

If you’re familiar with the pick/pickBy methods in Lodash, then this package should be easy to understand quickly. You can think of this package like a superset to pick from Lodash, with a feature rich schema syntax that allows picking properties from an object in a much more declarative way.

Yanking properties from objects can be done by providing the data you want to extract values from as the first parameter to the yank method, and the properties you wish to extract as either an array of schemas or as a series of arguments.

The schemas you provide to Yank follow a similar syntax to JSON objects, only with the values stripped. There are additional syntaxes for manipulating which values are picked called filters.

All syntax and demo examples below use tests/data.json as the source of data for picking values from, as shown below.

{
  "firstName": "John",
  "lastName": "Doe",
  "dateOfBirth": "1985-01-01",
  "addressDetails": {
    "address1": "10 Downing Street",
    "address2": null,
    "address3": null,
    "city": "London",
    "postcode": "SW1A 2AB"
  },
  "nested": {
    "data": {
      "items": {
        "one": "one",
        "two": "two",
        "three": "three"
      }
    }
  }
}

Syntax

Examples of schemas shown below are the strings you would provide as arguments to Yank, each followed by their resulting value. Each schema example is assumed to be the first and only schema parameter given to Yank.

Root property access

^Schema

firstName,
lastName

^Result

{
  "firstName": "John",
  "lastName": "Doe"
}

---

Nested properties

^Schema

addressDetails: {
  city
}

^Result

{
  "addressDetails": {
    "city": "London"
  }
}

---

Property path access

^Schema

addressDetails.city,
nested.data.items: {
  one,
  two
}

^Result

{
  "city": "London",
  "items": {
    "one": "one",
    "two": "two"
  }
}

---

Filters

On each key within a schema, filters can be applied to manipulate how Yank handles the value at that location. Filters are placed after the key, separated by a pipe (|), followed by parenthesis if the filter requires any arguments. If no arguments are required, the parenthesis can be omitted. Multiple filters can be applied simply by chaining additional filters separated by a pipe.

In addition, a macro flag can be applied to filters that support it by adding an exclamation point (!) to the end of the filter name.

Below is an example schema that makes use of filters.

^Schema

firstName | as(first_name),
lastName | as(last_name),
doesNotExist | nullable,
alsoDoesNotExist | nullable!: {
  thisWillBeNull,
  soWillThis
},
dateOfBirth | as(dob) | nullable,
thisWontBeInTheResult

^Result

{
  "first_name": "John",
  "last_name": "Doe",
  "doesNotExist": null,
  "alsoDoesNotExist": {
    "thisWillBeNull": null,
    "soWillThis": null
  },
  "dob": "1985-01-01"
}

Available filters

Name	Arguments	Macro	Description
as	(alias)		Renames the picked property
nullable		Sets child properties to nullable	Sets property to nullable
extract			Merges properties into the parent depth
exec	(key, …args)		Executes a method at the current depth with additional arguments passed to the method

Demo

Simply spread the property names into the yank method and it will return only those values from the given object.

^Usage

yank(data, 'firstName', 'lastName')

^Result

{
  "firstName": "John",
  "lastName": "Doe"
}

---

Alternatively, provide an array instead, or even a spread array of arrays, it’s all the same.

^Usage

yank(data, ['firstName'], ['lastName'])

^Result

{
  "firstName": "John",
  "lastName": "Doe"
}

---

All schemas are stripped of whitespace before parsing, so it’s easy and convenient to format them like you would for JSON.

^Usage

yank(data, `
  nested: {
    data: {
      items: {
        one,
        two
      }
    }
  }
`)

^Result

{
  "nested": {
    "data": {
      "items": {
        "one": "one",
        "two": "two"
      }
    }
  }
}

---

Have a nested object? Similar to JSON syntax, use { and } to yank child properties. This can be as nested as you like.

^Usage

yank(data, 'addressDetails: { address1, city }', 'firstName')

{
  "addressDetails": {
    "address1": "10 Downing Street",
    "city": "London"
  },
  "firstName": "John"
}

---

Deeply nesting is possible as well.

^Usage

yank(data, `
  nested: {
    data: {
      items: {
        one,
        two
      }
    }
  }
`)

{
  "nested": {
    "data": {
      "items": {
        "one": "one",
        "two": "two"
      }
    }
  }
}

Yanking properties that don’t exist will result in nothing happening for that particular key. The example below demonstrates that an empty object is returned because neither of the given properties exist on the original data object.

^Usage

yank(data, 'emailAddress', 'phoneNumber')

^Result

{}

---

Mixing existing properties with properties that don’t exist works too. The ones that don’t exist simply get ignored.

^Usage

yank(data, [
  `addressDetails: {
    county
  }`,
  `addressDetails: {
    city
  }`,
  'dateOfBirth'
])

^Result

{
  "addressDetails": {
    "city": "London"
  },
  "dateOfBirth": "1985-01-01"
}

---

Tests

$ npm test

Acknowledgments

• Heavily inspired by Lodash.pick and supick.