CircularJSON
Serializes and deserializes otherwise valid JSON objects containing circular references into and from a specialized JSON format.
Check out flatted too !!!
The future engine of CircularJSON is represented by flatted
- smaller output, compared to every other alternative
- fast, and usually much faster, than CircularJSON, on IE/Edge browser too
- smaller in size, easier to maintain, 30 LOC proof of concept
The next release of CircularJSON will be based on flatted
representation which is,
as the term indicates, a flat array of references.
That means that produced and imported strings will not be compatible with CircularJSON one.
If you care about improving serialization of your data, and you don't need extra JSON API arguments,
try to see if flatted
works already out of the box.
A Working Solution To A Common Problem
A usage example:
var object = {};
object.arr = [
object, object
];
object.arr.push(object.arr);
object.obj = object;
var serialized = CircularJSON.stringify(object);
var unserialized = CircularJSON.parse(serialized);
unserialized.obj === unserialized;
unserialized.arr[0] === unserialized;
unserialized.arr.pop() === unserialized.arr;
A quick summary:
- new in version
0.5
, you can specify a JSON parser different from JSON itself. CircularJSON.parser = ABetterJSON;
is all you need. - uses
~
as a special prefix symbol to denote which parent the reference belongs to (i.e. ~root~child1~child2
) - reasonably fast in both serialization and deserialization
- compact serialization for easier and slimmer transportation across environments
- tested and covered over nasty structures too
- compatible with all JavaScript engines
Node Installation & Usage
npm install --save circular-json
'use strict';
var
CircularJSON = require('circular-json'),
obj = { foo: 'bar' },
str
;
obj.self = obj;
str = CircularJSON.stringify(obj);
There are no dependencies.
Browser Installation & Usage
- Global: <build/circular-json.js>
- AMD: <build/circular-json.amd.js>
- CommonJS: <build/circular-json.node.js>
(generated via gitstrap)
<script src="build/circular-json.js"></script>
'use strict';
var CircularJSON = window.CircularJSON
, obj = { foo: 'bar' }
, str
;
obj.self = obj;
str = CircularJSON.stringify(obj);
NOTE: Platforms without native JSON (i.e. MSIE <= 8) requires json3.js
or similar.
It is also a bad idea to CircularJSON.parse(JSON.stringify(object))
because of those manipulation used in CircularJSON.stringify()
able to make parsing safe and secure.
As summary: CircularJSON.parse(CircularJSON.stringify(object))
is the way to go, same is for JSON.parse(JSON.stringify(object))
.
API
It's the same as native JSON, except the fourth parameter placeholder
, which circular references to be replaced with "[Circular]"
(i.e. for logging).
- CircularJSON.stringify(object, replacer, spacer, placeholder)
- CircularJSON.parse(string, reviver)
Bear in mind JSON.parse(CircularJSON.stringify(object))
will work but not produce the expected output.
Similar Libraries
Why Not the @izs One
The module json-stringify-safe seems to be for console.log()
but it's completely pointless for JSON.parse()
, being latter one unable to retrieve back the initial structure. Here an example:
{
"circularRef": "[Circular]",
"list": [
"[Circular]",
"[Circular]"
]
}
Just type this in your node
console: var o = {}; o.a = o; console.log(o);
. The output will be { a: [Circular] }
... good, but that ain't really solving the problem.
However, if that's all you need, the function used to create that kind of output is probably faster than CircularJSON
and surely fits in less lines of code.
Why Not {{put random name}} Solution
So here the thing: circular references can be wrong but, if there is a need for them, any attempt to ignore them or remove them can be considered just a failure.
Not because the method is bad or it's not working, simply because the circular info, the one we needed and used in the first place, is lost!
In this case, CircularJSON
does even more than just solve circular and recursions: it maps all same objects so that less memory is used as well on deserialization as less bandwidth too!
It's able to redefine those references back later on so the way we store is the way we retrieve and in a reasonably performant way, also trusting the snappy and native JSON
methods to iterate.