API reference

Reductions with where= / out=

All seven scalar reductions (sum, prod, min, max, mean, var, std) accept where (a boolean mask broadcast to a) and out (a pre-allocated result array). Only elements where the mask is true contribute; masked-out positions use the reduction identity (0 for sum, 1 for prod; for min/max you must supply initial). mean/var/std divide by the count of true positions.

Name	Summary
`sum`	`where` is a bool `ArrayLike` broadcast to `a`.
`mean`	Mean of the unmasked elements; divisor is the count of `true` positions.
`var`	Variance / standard deviation over unmasked elements.
`min`	`min`/`max` with a mask require `initial` (the identity / seed value), matching NumPy's requirement.
`out=`	Write the result into a pre-allocated `NDArray`.

`sum (where=)`

np.sum(a, { where, out? })

where is a bool ArrayLike broadcast to a. Masked-out positions contribute 0.

NumPynumpy.sum

Parameters

aArrayLike: An NDArray, nested JS array or scalar.
options.whereArrayLike: Boolean mask; true positions are summed.
[options.out]NDArray: Pre-allocated output (exact result shape).

Returns

NDArray

Example

TypeScript

const a = np.array([1, 2, 3, 4, 5], { dtype: "float64" });
const mask = np.array([true, false, true, false, true]);
np.sum(a, { where: mask }).item();   // => 9

TypeScript declaration

np.sum(a: ArrayLike, opts?: ReduceOptions | undefined): NDArray

`mean (where=)`

np.mean(a, { where, out? })

Mean of the unmasked elements; divisor is the count of true positions.

NumPynumpy.mean

Parameters

aArrayLike: An NDArray, nested JS array or scalar.
options.whereArrayLike: Boolean mask.

Returns

NDArray

Example

TypeScript

const a = np.array([1, 2, 3, 4, 5], { dtype: "float64" });
const mask = np.array([true, false, true, false, true]);
np.mean(a, { where: mask }).item();  // => 3

TypeScript declaration

np.mean(a: ArrayLike, opts?: Omit<ReduceOptions, "initial"> | undefined): NDArray

`var / std (where=)`

np.var(a, { where, ddof? }) · np.std(a, { where, ddof? })

Variance / standard deviation over unmasked elements.

NumPynumpy.var numpy.std

Parameters

aArrayLike: An NDArray, nested JS array or scalar.
options.whereArrayLike: Boolean mask.
[options.ddof]number: Delta degrees of freedom (default 0).

Returns

NDArray

Example

TypeScript

const a = np.array([1, 2, 3, 4, 5], { dtype: "float64" });
const mask = np.array([true, false, true, false, true]);
np.var(a, { where: mask }).item();   // => 2.6666666666666665
np.std(a, { where: mask }).item();   // => 1.632993161855452

TypeScript declaration

np.var(a: ArrayLike, opts?: VarOptions | undefined): NDArray
np.std(a: ArrayLike, opts?: VarOptions | undefined): NDArray

`min / max (where= + initial=)`

np.min(a, { where, initial }) · np.max(a, { where, initial })

min/max with a mask require initial (the identity / seed value), matching NumPy's requirement.

NumPynumpy.min numpy.max

Parameters

aArrayLike: An NDArray, nested JS array or scalar.
options.whereArrayLike: Boolean mask.
options.initialnumber: Required seed value.

Returns

NDArray

Example

TypeScript

const a = np.array([1, 2, 3, 4], { dtype: "float64" });
const mask = np.array([false, true, false, true]);
np.min(a, { where: mask, initial: 1e10 }).item();  // => 2
np.max(a, { where: mask, initial: -1e10 }).item(); // => 4

TypeScript declaration

np.min(a: ArrayLike, opts?: Omit<ReduceOptions, "dtype"> | undefined): NDArray
np.max(a: ArrayLike, opts?: Omit<ReduceOptions, "dtype"> | undefined): NDArray

`out= parameter`

np.sum/prod/min/max/mean/var/std(a, { out })

Write the result into a pre-allocated NDArray. out.shape must exactly match the result shape. The same array is returned.

NumPynumpy.sum

Parameters

aArrayLike: An NDArray, nested JS array or scalar.
options.outNDArray: Pre-allocated output (exact result shape).

Returns

NDArray (same object as out)

Example

TypeScript

const a = np.array([1, 2, 3, 4, 5], { dtype: "float64" });
const out = np.zeros([]);
np.sum(a, { out });
out.item();  // => 15

TypeScript declaration

np.sum(a: ArrayLike, opts?: ReduceOptions | undefined): NDArray