Awesome

ndarray-concat-rows

Concatenate ndarrays by row (along the first dimension)

Introduction

This module takes a list of input ndarrays and concatenates it along the first dimension. That is, a 3 × 2 ndarray concatenated with a 10 × 2 ndarray yields a 13 × 2 ndarray.

Examples

Understanding (+) in the comments below to indicate row concatenation,

var ndarray = require('ndarray')
var r = require('ndarray-concat-rows')

// Concatenating vectors:
//   [1 2] (+) [3 4] -> [1 2 3 4]
//
r([ ndarray([1, 2]), ndarray([3, 4]) ])
// => ndarray([1, 2, 3, 4])

// Concatenating matrices:
//                       [1  2]
//   [1 2]     [5  6]    [3  4]
//   [3 4] (+) [7  8] -> [4  5]
//             [9 10]    [7  8]
//                       [9 10]
//
r([ ndarray([1, 2, 3, 4], [2, 2]), ndarray([5, 6, 7, 8, 9, 10], [3, 2]) ])
// => ndarray([1, 2, 3, 4, 5, 6, 7, 8, 9, 10], [5, 2])

Installation

$ npm install ndarray-concat-rows

API

`require('ndarray-concat-rows')([output,] input, [options])`

Arguments:

output (optional): An optional output destination. The shape must match the shape of the concatenated arrays, otherwise an error will be thrown. If not provided, storage will be allocated using ndarray-scratch.
input: A javascript Array containing ndarrays to be concatenated. If this is missing or empty, an error will be thrown. Given n-dimensional input, all arguments must have the same dimensionality and the last n-1 dimensions of each arguments must have the same length.
options (optional): An optional object containing options. Options are:
- dtype: If no output ndarray is provided, the dtype of the output will be double (equivalently float64) by default, or otherwise the dtype specified here. See ndarray dtypes.

Returns: A reference to the output ndarray containing the concatenated data.