The ultimate JSON parser
Piffero is an open source SAX-like parser who work directely on the streams to get parts of big JSON files.
Piffero can load big files larger than memory used and return the required content in a stream.
______________________________ . ♪
=(____|_o_o_o_o_|_o_o_o_o_|*____| | ♫
'
The version 1.0.0 only support the full qualified JSONpath
example: $.father.son.array[:index].attribute
The sintax that will be implemented in the next future and the examples from Stefan Goessner's post
JSONPath | Description | Implemented |
---|---|---|
$ |
The root object/element | ✅ |
@ |
The current object/element | |
. |
Child member operator | ✅ |
.. |
Recursive descendant operator; JSONPath borrows this syntax from E4X | |
* |
Wildcard matching all objects/elements regardless their names | |
[] |
Subscript operator | ✅ |
[,] |
Union operator for alternate names or array indices as a set | ✅ |
[start:end:step] |
Array slice operator borrowed from ES4 / Python | ✅ |
?() |
Applies a filter (script) expression via static evaluation | |
() |
Script expression via static evaluation |
See the replit.com demo.
Step 1: install Piffero
npm i piffero
Step 2: use Piffero
Piffero is easy to use, you only need to call one of the methods:
findByPath(inputStream: Stream, jsonpath: string): Stream
or
findAsString(callback: (result: any, err?: any) => void, inputStream: Stream, jsonpath: string): void
or
findAsPromise(stream: Readable, jsonPath: string): Promise<string>
Below there are some examples of use case based on this json file:
{
"employees":[
{
"name":"John",
"surname":"Doe",
"phoneNumbers":[
{
"type":"iPhone",
"number":"0123-4567-8888",
"test":false
},
{
"type":"home",
"number":"0123-4567-8910",
"test":true
}
]
},
{
"name":"Joe",
"surname":"Black",
"phoneNumbers":[
]
}
]
}
For example, if you want extract the second employees "Joe Black", you need to call findByPath
with the json path $.employees[1]
, eg:
const piffero = require('piffero');
const fs = require('fs');
// create read stream of json
const inputStream = fs.createReadStream('./employees.json');
// pass the stream to Piffero with the json path
const resultStream = piffero.Piffero.findByPath(inputStream, "$.employees[1]");
// trasform the stream to string and print into console
const resultString = '';
resultStream.on('data', chunk => resultString += chunk.toString());
resultStream.on('end', () => console.log(resultString));
[{"name":"Joe","surname":"Black","phoneNumbers":[]}]
Otherwise, if you want the second phone number of the first employees, you need to call findByPath
with the json path $.employees[0].phoneNumbers[1]
, eg:
const piffero = require('piffero');
const fs = require('fs');
// create read stream of json
const inputStream = fs.createReadStream('./employees.json');
// pass the stream to Piffero with the json path
piffero.Piffero.findAsString((result, err) => console.log(result), inputStream, '$.employees[0].phoneNumbers[1]');
[{"type":"home","number":"0123-4567-8910","test":true}]
import * as fs from "fs";
import * as express from "express";
import { Piffero } from "piffero";
const app: express.Application = express();
app.get("/api/:jsonPath", (req, res) => {
const jsonPath: string = req.params.jsonPath;
const inputStream: fs.ReadStream = fs.createReadStream('./employees.json');
const resultStream = Piffero.findByPath(inputStream, jsonPath);
res.setHeader("Content-Type", "application/json");
resultStream.pipe(res);
});
app.listen(3000, () => {
console.log("Open your browser at: http://localhost:3000/api/$.employees[0].phoneNumbers[1]");
});
benchmarck tests was done to compare Piffero with Oboe
using benchmark
Hardware: i7-4510U 12GB RAM
Parser | ops/secc | runs sampled |
---|---|---|
Piffero (stream) | 1.0186262222075857 ops/sec | 10 |
Piffero (string) | 1.1235041104521386 ops/sec | 10 |
Oboe | 0.4287069078285109 ops/sec | 7 |
Parser | ops/secc | runs sampled |
---|---|---|
Piffero (stream) | 2.192788691917704 ops/sec | 17 |
Piffero (string) | 2.4804025718178084 ops/sec | 16 |
Oboe | 0.5590235602555842 ops/sec | 7 |
Parser | ops/secc | runs sampled |
---|---|---|
Piffero (stream) | 206.17873294205398 ops/sec | 68 |
Piffero (string) | 652.333851058612 ops/sec | 42 |
Oboe | 7.3452430736806225 ops/sec | 7 |
Piffero is almost 3 time faster than Oboe if both have to stream the whole file, the difference becomes bigger if they don't have to steam the whole file.
- Piffero is built on Clarinet
- You can try Oboe
- You can also try JSONPath Online Evaluator
enjoy Piffero!!