NumPy · Pipe

ndarray.size

ndarray.size gives the total number of elements in an array. It equals the product of the array’s dimensions and returns a standard arbitrary-precision Python integer.

How it works

Accessing an array’s size returns the total number of elements across all dimensions. It is equivalent to multiplying the dimensions in shape, but size returns a standard arbitrary-precision Python integer.

Syntax

a.size

Key points

The attribute reports the array’s complete element count.
Its value is the product of every dimension listed in shape.
The result is a standard Python integer with arbitrary precision.
Other ways to compute the same count may use a different integer type.
The documented np.prod(a.shape) alternative returns np.int_.
The page provides an executable example demonstrating the attribute.

Common mistakes

Do not assume that equivalent counting expressions return the same integer type: size returns a standard Python integer, while some alternatives may return a NumPy integer.
Prefer size when the count will participate in calculations where fixed-width integer overflow is a concern.

Example: Count every element

import numpy as np

array = np.zeros((3, 5, 2), dtype=np.complex128)
print(array.size)  # 30
print(np.prod(array.shape))  # 30

Creates a three-dimensional array and reads its total element count with the size attribute.

When to use it

Shape and size checks

Validate expected tensor or matrix dimensions before processing.

Array-size-driven workflows

Allocate, reshape, or report data using the total element count.

Safe count arithmetic

Avoid fixed-width integer concerns when using counts in later calculations.

Interview questions

QWhat does ndarray.size return?

ndarray.size is an attribute containing the total number of elements in the array. For an array with shape (3, 5, 2), its value is 30.

Follow-up: How can the total be derived from shape? Multiply all dimensions together; this is equivalent to np.prod(a.shape).

QHow does size relate to shape?

The total is the product of the array’s dimensions, so size equals np.prod(a.shape). It counts elements across every axis rather than along only one axis.

Follow-up: How would you count elements along one axis instead? Use the documented np.size function with its axis argument, rather than ndarray.size.

QWhat type does ndarray.size produce?

a.size returns a standard arbitrary-precision Python integer. An alternative such as np.prod(a.shape) returns a NumPy integer type, so their types can differ even when their numeric values match.

Follow-up: Why can that type difference matter? A fixed-size integer result can be vulnerable to overflow in later calculations, whereas the standard Python integer avoids that limitation.

numpy.ndarray.shape
numpy.prod
numpy.size

Source: numpy.ndarray.size — NumPy v2.5 Manual. Checked 2026-10-15.