Skip to content

Latest commit

 

History

History
109 lines (85 loc) · 2.62 KB

README.md

File metadata and controls

109 lines (85 loc) · 2.62 KB

pandas-js

Pandas for JavaScript

pandas.js is an open source (experimental) library mimicking the Python pandas library. It relies on Immutable.js as the NumPy base. The main data objects in pandas.js are the Series and the DataFrame

Documentation

See the docs
See also this post on use for optimizing React logic.

Installation and use

$ npm install pandas-js

Importing Series and DataFrame

import { Series, DataFrame } from 'pandas-js';

Create a new Series

const ds = new Series([1, 2, 3, 4], {name: 'My test name', index: [2, 3, 4, 5]})
ds.toString()
// Returns:
// 2  1
// 3  2
// 4  3
// 5  4
// Name: My test name, dtype: dtype(int)

Filter a Series

const ds = new Series([1, 2, 3]);

// Returns Series([2, 3]);
ds.filter(ds.gte(2));

Filtering can be done with generic methods

const ds = new Series([1, 2, 3], {name: 'Test name'})

// Returns Series([true, false, false])
ds.eq(1);

// Returns Series([false, true, true])
ds.eq(new Series([0, 2, 3]));

// Returns Series([false, true, true])
ds.eq(Immutable.List([0, 2, 3]));

// Returns Series([false, true, true])
ds.eq([0, 2, 3]);

// Returns Series([2, 3])
ds.filter(ds.eq([0, 2, 3]));

Create a new DataFrame

const df = new DataFrame([{x: 1, y: 2}, {x: 2, y: 3}, {x: 3, y: 4}])

// Returns:
//    x  |  y
// 0  1  |  2
// 1  2  |  3
// 2  3  |  4
df.toString();

Filtering a DataFrame

const df = new DataFrame(Immutable.Map({x: new Series([1, 2]), y: new Series([2, 3])}));

// Returns DataFrame(Immutable.Map({x: Series([2]), y: Series([3]));
df.filter(df.get('x').gt(1));

// Returns DataFrame(Immutable.Map({x: Series([2]), y: Series([3]));
df.filter([false, true]);

// Returns DataFrame(Immutable.Map({x: Series([2]), y: Series([3]));
df.filter(Immutable.Map([false, true]));

Development

Testing and build

$ npm run test
$ npm run build

Testing uses Jest. Building requires the babel compiler.