Combine-JSON

This module (and CLI) allows you to create one JSON file from smaller files stored in a directory hierarchy. The smaller files can be in JSON (default) or any other format like JSON5 or YAML or even INI (custom parser).

Why? Sometimes you have a huge JSON file with lots of nested objects but you want to break it into a file hierarchy because:

• It is easier to browse to a particular section
• It is easier to understand the shape of the data structure at a glance
• Editing any subparts of the document leads to cleaner individual git history and diffs

Example

Given this JSON file:

my-data.json

{
    "name": "Alex Ewerlöf",
        "address": {
        "street": "Hittepågatan 13",
        "city": "Stockholm",
        "country": "Sweden",
        "zip": "11122"
    },
    "todos": [
        {
            "id": 1,
            "title": "document the module",
        },
        {
            "id": 2,
            "title": "write some tests",
        },
        {
            "id": 3,
            "title": "publish it for good",
        }
    ]
}

We can break it into several files like this:

my-data/name.json

"Alex Ewerlöf"

(Yes that's a perfectly valid input to JSON.parse())

my-data/address.json

{
    "street": "Hittepågatan 13",
    "city": "Stockholm",
    "country": "Sweden",
    "zip": "11122"
}

my-data/todos.json

[
    {
        "id": 1,
        "title": "document the module",
    },
    {
        "id": 2,
        "title": "write some tests",
    },
    {
        "id": 3,
        "title": "publish it for good",
    }
]

You can write that array of objects into files as well:

my-data/todos/0.json

{
    "id": 1,
    "title": "document the module",
}

my-data/todos/1.json

{
    "id": 2,
    "title": "write some tests",
}

my-data/todos/2.json

{
    "id": 3,
    "title": "publish it for good",
}

So the file structure looks like this:

my-data/
    |____name.json
    |____address.json
    |____todos/
        |____0.json
        |____1.json
        |____2.json

Of course you can have directories for array elements as well:

my-data/
    |____name.json
    |____address.json
    |____todos/
        |____0/
        |    |____id.json
        |    |____title.json
        |____1.json
        |____2.json

Take a look at the test/my-data directory to see it in action.

Usage

$ npm i combine-json

const { combine } = require('combine-json')

const myBigJsonObject = await combine('path/to/the/root/dir')

API

See the js docs online.

Known limitations

• In the current implementation we ignore any directory starting with ..
• If a folder contains subfolders or files that look like numbers, an array will be created instead of an object.

CLI

You can use the CLI for testing what the output may look like.

$ npx combine-json test/my-data/
{
    "address": {
        "street": "Hittepågatan 13",
        "city": "Stockholm",
        "country": "Sweden",
        "zip": "11122"
    },
    "name": "Alex Ewerlöf",
    "todos": [
        {
            "id": 1,
            "title": "document the module"
        },
        {
            "id": 2,
            "title": "write some tests"
        },
        {
            "id": 3,
            "title": "publish it for good"
        }
    ]
}

Made in Sweden by @alexewerlof