map

Applies a transformation function to each element and returns a new array of the same shape. Indexed variants provide the element's flat index (mapIndexed, 1D only) or multi-dimensional index (mapMultiIndexed). The result element type R must be a supported DataType.

Signatures

inline fun <T, D : Dimension, reified R : Any> MultiArray<T, D>.map(
    transform: (T) -> R
): NDArray<R, D>

inline fun <T, reified R : Any> MultiArray<T, D1>.mapIndexed(
    transform: (index: Int, T) -> R
): D1Array<R>

inline fun <T, D : Dimension, reified R : Any> MultiArray<T, D>.mapMultiIndexed(
    transform: (index: IntArray, T) -> R
): NDArray<R, D>

inline fun <T, D : Dimension, reified R : Any> MultiArray<T, D>.mapNotNull(
    transform: (T) -> R?
): NDArray<R, D>

Parameters

Parameter	Type	Description
`transform`	`(T) -> R`	Maps each element to a new value.
`transform`	`(index: Int, T) -> R`	With flat index (1D arrays only).
`transform`	`(index: IntArray, T) -> R`	With multi-dimensional index.

Returns: NDArray<R, D> — a new array with the same shape and transformed elements.

Example

val a = mk.ndarray(mk[mk[1, 2], mk[3, 4]])

val doubled = a.map { it * 2 }
// [[2, 4],
//  [6, 8]]

val b = mk.ndarray(mk[10, 20, 30])
b.mapIndexed { i, v -> v + i }
// [10, 21, 32]

a.mapMultiIndexed { idx, v -> idx.sum() + v }
// [[1, 3],
//  [4, 6]]

28 February 2026