Documentation v1.2.1
Home GitHub

Serialization

This page explains how the node tree is written back out as JSON: the node-level json getter, the root-level editor.json and editor.export(), and how the two root paths differ.

Node Level: JSONEditorNode.json

JSONEditorNode.json (a getter) recursively calls the private #json() (JSONEditorNode.js:157), returning a JavaScript value per node type:

Node Type Returns
array An array of each child's #json() result
object An object keyed by each child's key
boolean value.toLowerCase() === "true"
number Number(value)
anything else The raw value

When an object node builds its object, keyless children are skipped — except when that object node itself sits inside an array (this.parent.type === "array"), in which case a keyless child gets the current count of collected keys as a numeric key (JSONEditorNode.js:170). For duplicate keys, the later one wins.

Root Level: editor.json vs. editor.export()

The two root-level paths filter differently:

Path Filter Condition (JSONEditor.js) Output
editor.json (getter, lines 267-277) Only children where e.key is truthy A 4-space-indented JSON string
editor.export() (lines 289-310) Children where e.key is truthy, or any child when children.length === 1 Downloads NanoJSON-{timestamp}.json, also 4-space-indented

Both paths build an object as the root; a keyless child is written under the key 0 (result[e.key || 0]).

Importing an Array

The root's editor.type is always "object". When the imported data is an array, the root-level children are keyless, so the two paths diverge:

Imported Data editor.json editor.export() Download
[1] "{}" {"0": 1}
[1, 2] "{}" {}
{"a": 1} {"a": 1} {"a": 1}

A root-level array is never written out as an array. To keep an array root, wrap it in an object before passing it in (for example { "items": [...] }).

Further Reading

中文