bjoerge / json-reduce

Reduce any JSON value by visiting each node depth first

json-reduce

Reduce any JSON value by traversing depth first and visiting each node

Example: calculate the sum of all number values in a json object

import {reduce} from 'json-reduce'

const document = {
  first: 1,
  second: 2,
  deep: {
    array: [3, 4, 5, 6],
    seven: 7
  }
}

const result = reduce(
  document,
  (acc, value, path) => (typeof value === 'number' ? acc + value : acc),
  0
)

console.log(result)
//=> 28

API

reduce(value, reducer, initialValue)

reducer is the reducer function to execute for each node in the tree, and is given three arguments:

• accumulator - The accumulation of the callback's return values; it is the value returned from the previous invocation of the callback, or initialValue.
• value - The current node being traversed
• path - The "dot-path" to the current node being traversed, e.g. ['deep', 'array', 2]

Skipping subtrees

Sometimes when encountering a specific object or array value, you want to skip traversing the subtree. This can be done calling a provided SKIP function like this:

import reduce, {SKIP} from 'json-reduce'

const doc = {
  species: [
    {name: 'clover', type: 'plant'},
    {name: 'trout', type: 'fish', eats: [{type: 'animal', name: 'crayfish'}]},
    {
      type: 'animal',
      name: 'bear',
      food: [
        {
          type: 'animal',
          name: 'deer',
          food: [{type: 'plant', name: 'leaves'}]
        },
        {
          type: 'plant',
          name: 'blueberry'
        }
      ]
    }
  ]
}

const result = reduce(
  doc,
  (acc, val, path) => {
    if (val.type === 'plant' || val.type === 'fish') {
      // We don't want to traverse the subtrees of these
      return SKIP
    }
    // Collect all traversed paths
    return acc.concat([path])
  },
  []
)
expect(result).toEqual([
  [],
  ['species'],
  ['species', 2],
  ['species', 2, 'type'],
  ['species', 2, 'name'],
  ['species', 2, 'food'],
  ['species', 2, 'food', 0],
  ['species', 2, 'food', 0, 'type'],
  ['species', 2, 'food', 0, 'name'],
  ['species', 2, 'food', 0, 'food']
])

Return and skip

In addition to return SKIP, you can also call SKIP with a return value for convenience, to both return the accumulated value and signal subtree skipping in one operation, e.g.:

//
reduce(doc, (acc, node) => {
  if (node.type === 'plant' || node.type === 'fish') {
    // Uppercase plant and fish names, but skip traversing subtrees
    return SKIP(
      acc.concat({
        ...node,
        name: node.name.toUpperCase()
      })
    )
  }
  return acc
}, [])

Gotchas / Limitations

• Initial value is required. json-reduce does not work like Array.prototype.reduce with respect to missing initial value
• No circular reference detection and handling. Passing a circular structure to reduce() will probably crash with maximum call stack size exceeded.
• Assumes a data structure that consists of valid JSON data types only. Traversal of Map, Set, etc. are not currently supported (PR welcome!).