Datamodel Save Abandoned
Relational algebra compliant in memory tabular data store.
What is DataModel?
DataModel is an in-browser representation of tabular data. It uses WebAssembly for high performance and works seamlessly with any JavaScript library. It supports Relational Algebra operators which enable you to run select, group, sort (and many more) operations on the data.
The current version performs all the data operations like filtering, aggregation, etc. on WebAssembly which gives a 10x performance boost compared to the old JavaScript version.
It is written in Rust Language to handle computation intensive data operations, which is then compiled to WebAssembly, thereby providing a native-like performance for data operations.
DataModel can be used if you need an in-browser tabular data store for data analysis, visualization or just general use of data.
Features
-
? Supports Relational Algebra operators e.g.
selection,projection,group,calculateVariable,sortetc out-of-the-box. -
? Every operation creates Immutable DataModel instance and builds a Directed Acyclic Graph (DAG) which establishes auto interactivity.
-
? Uses WebAssembly for handling huge datasets and for better performance.
-
⛺ Also works in Nodejs environment out-of-the-box.
Installation
CDN
Insert the DataModel build into the <head>:
<script src="https://cdn.jsdelivr.net/npm/@chartshq/[email protected]/dist/browser/datamodel.js" type="text/javascript"></script>
NPM
Install DataModel from NPM:
$ npm install --save @chartshq/datamodel
As we're using Worker internally, so the worker-loader needs to be installed as follows:
$ npm install worker-loader --save-dev
And then within your webpack configuration object, you'll need to add the worker-loader to the list of module rules, like so:
module.exports = {
module: {
rules: [
// Add the following object to your module `rules` list.
{
test: /\.worker/,
include: /datamodel/,
loader: 'worker-loader',
options: {
inline: false, // If you want to make it inline, set to true.
fallback: true
},
},
],
}
};
You also can checkout our datamodel-app-template to try out the DataModel quickly through a boilerplate app.
Getting Started
Once the installation is done, please follow the steps below:
- Prepare the data and the corresponding schema:
// Prepare the schema for data.
const schema = [
{
name: 'Name',
type: 'dimension'
},
{
name: 'Maker',
type: 'dimension'
},
{
name: 'Horsepower',
type: 'measure',
defAggFn: 'avg'
},
{
name: 'Origin',
type: 'dimension'
}
]
// Prepare the data.
const data = [
{
"Name": "chevrolet chevelle malibu",
"Maker": "chevrolet",
"Horsepower": 130,
"Origin": "USA"
},
{
"Name": "buick skylark 320",
"Maker": "buick",
"Horsepower": 165,
"Origin": "USA"
},
{
"Name": "datsun pl510",
"Maker": "datsun",
"Horsepower": 88,
"Origin": "Japan"
}
]
- Import DataModel as follows:
If you are using the npm package, import the package as below:
import Engine from '@chartshq/datamodel';
If you are using it in NodeJS, then require it as below:
const Engine = require('@chartshq/datamodel').default;
If you are using CDN, then use it as follows:
const Engine = window.DataModel;
- Load the DataModel engine and pass the data and schema to
DataModelconstructor and create a newDataModelinstance:
// As the DataModel are asynchronous, so we need to
// use async-await syntax.
async function myAsyncFn() {
// Load the DataModel module.
const DataModel = await Engine.onReady();
// Converts the raw data into a format
// which DataModel can consume.
const formattedData = await DataModel.loadData(data, schema);
// Create a new DataModel instance with
// the formatted data.
const dm = new DataModel(formattedData);
console.log(dm.getData().data);
// Output:
// [
// ["chevrolet chevelle malibu", "chevrolet", 130, "USA"],
// ["buick skylark 320", "buick", 165, "USA"],
// ["datsun pl510", "datsun", 88, "Japan"]
// ]
// Perform the selection operation.
const selectDm = dm.select({ field: 'Origin', value: 'USA', operator: DataModel.ComparisonOperators.EQUAL });
console.log(selectDm.getData().data);
// Output:
// [
// ["chevrolet chevelle malibu", "chevrolet", 130, "USA],
// ["buick skylark 320", "buick", 165, "USA]
// ]
// Perform the projection operation.
const projectDm = dm.project(["Origin", "Maker"]);
console.log(projectDm.getData().data);
// Output:
// [
// ["USA", "chevrolet"],
// ["USA", "buick"],
// ["Japan", "datsun"]
// ]
console.log(projectDm.getData().schema);
// Output:
// [
// {"name": "Origin","type": "dimension"},
// {"name": "Maker","type": "dimension"}
// ]
}
myAsyncFn()
.catch(console.error.bind(console));
- Now dispose the DataModel instance if it's not needed:
// This also disposes all the datamodels which are created from it.
dm.dispose();
Documentation
Find detailed documentation and API reference from here.
What has changed?
DataModel 3.0.0 now has the core written in Rust language and has been ported to WebAssembly bringing in a huge performance difference w.r.t to previous version, in terms of both data size and computing speed. While the JavaScript version is deprecated and no active development will take place there but critical bugs if raised would be taken and released in GitHub only.
You can visit the JavaScript (deprecated) version here https://github.com/chartshq/datamodel-deprecated
Migrating from previous versions of DataModel
Now the DataModel became asynchronous as opposed to being synchronous in the previous JavaScript version.
import Engine from '@chartshq/datamodel';
(async () => {
// Load the DataModel module.
const DataModel = await Engine.onReady();
// Converts the raw data into a format
// which DataModel can consume.
const formattedData = await DataModel.loadData(data, schema);
// Create a new DataModel instance with
// the formatted data.
const dm = new DataModel(formattedData);
})();
Changed APIs
-
select
DataModel deprecated version:
dm.select((fields) => { return fields.Origin.value === 'USA'; });Latest version:
dm.select({ field: 'Origin', operator: DataModel.ComparisonOperators.EQUAL, value: 'USA' }); -
groupBy
DataModel deprecated version:
dm.groupBy(['Origin'], { Acceleration: 'avg' });Latest version:
dm.groupBy(['Origin'], [{ aggn: DataModel.AggregationFunctions.AVG, field: 'Acceleration' }]);
Supported data operations:
- select
- project
- calculateVariable
- sort
- groupBy
Upcoming data operations:
- join
- bin
- compose
- union
- difference
- ... many more ...
For more details on APIs visit our docs.
License
Custom License (Free to use)
Open Source Agenda Badge
