ndarray

Overview

mk.ndarray() is the most common way to create arrays. It infers type and dimension from the input. Use mk[...] to build nested list literals inline.

From nested lists with mk[]

val v = mk.ndarray(mk[1, 2, 3])                         // D1Array<Int>
val m = mk.ndarray(mk[mk[1.0, 2.0], mk[3.0, 4.0]])     // D2Array<Double>
val t = mk.ndarray(mk[mk[mk[1, 2], mk[3, 4]]])         // D3Array<Int>

Overloads exist for 1D–4D with both Number and Complex element types.

From collection with shape

val a = mk.ndarray<Int, D2>(listOf(1, 2, 3, 4), intArrayOf(2, 2))
// D2Array<Int>, shape (2, 2)

Parameter	Type	Description
`elements`	`Collection<T>`	Flat collection of values.
`shape`	`IntArray`	Target shape. Product must equal `elements.size`.

Dimension can be reified or passed explicitly:

val b = mk.ndarray(listOf(1, 2, 3, 4), intArrayOf(2, 2), D2)

From collection with dimension sizes

val m = mk.ndarray(listOf(1, 2, 3, 4, 5, 6), 2, 3)     // D2Array, shape (2, 3)
val t = mk.ndarray(listOf(1, 2, 3, 4, 5, 6), 1, 2, 3)  // D3Array, shape (1, 2, 3)

Overloads for 2D, 3D, 4D, and N-D (5+ dimensions via varargs).

From primitive arrays

1D:

val a = mk.ndarray(intArrayOf(1, 2, 3))         // D1Array<Int>
val b = mk.ndarray(doubleArrayOf(1.0, 2.0))     // D1Array<Double>

With shape:

val m = mk.ndarray(intArrayOf(1, 2, 3, 4), 2, 2) // D2Array<Int>

Supported types: ByteArray, ShortArray, IntArray, LongArray, FloatArray, DoubleArray, ComplexFloatArray, ComplexDoubleArray.

From Array of primitive arrays

val m = mk.ndarray(arrayOf(intArrayOf(1, 2), intArrayOf(3, 4)))  // D2Array<Int>

Pitfalls

The product of the shape must exactly equal the number of elements. A mismatch throws IllegalArgumentException.
Nested list inputs are validated for consistent sizes — jagged lists are rejected.

28 February 2026