Track compounded values
Compute running totals of multiplicative quantities such as growth factors or conversion rates.
NumPy · Function
np.cumprod
Returns running products of array elements along a selected axis. Without axis, it flattens the input; dtype controls multiplication and output types, and out can reuse an existing array.
How it works
np.cumprod returns running products along a chosen axis. With no axis, it flattens the input first; dtype controls the accumulator and result type, while out can receive the result in a supplied array.
Syntax
numpy.cumprod(a, axis=None, dtype=None, out=None)Parameters
| Name | Type | Description |
|---|---|---|
a | array_like | Input array. |
axis | int, optional | Axis along which the cumulative product is computed. If omitted, the input is flattened. |
dtype | dtype, optional | Type of the returned array and the accumulator used for multiplication. |
out | ndarray, optional | Existing output array for the result; its shape and buffer length must match the expected output. |
Returns
An ndarray containing the cumulative products, with the shape determined by the input and axis; if out is supplied, the result is written there.
Key points
- The function produces prefix multiplications rather than one final product.
- Use axis to select the dimension over which running multiplication occurs.
- dtype controls both the output array and the intermediate multiplication accumulator.
- Use out when the result should be written into an existing ndarray.
- For a matrix, axis=0 accumulates independently down columns.
- For a matrix, axis=1 accumulates independently across rows.
Common mistakes
- Leaving axis unset changes a multidimensional array into a one-dimensional traversal before multiplication.
- An out array must match the expected result shape and buffer size.
Example: Cumulative products by row
import numpy as np
values = np.array([[1, 2, 3], [4, 5, 6]])
print(np.cumprod(values, axis=1))
# [[ 1 2 6]
# [ 4 20 120]]Computes running products independently across each row.
When to use it
Accumulate along matrix dimensions
Generate prefix products independently across rows or columns of a matrix.
Manage dtype and memory
Control numeric representation or write into preallocated storage.
Interview questions
QWhat does np.cumprod compute, and what is the default axis behavior?
It computes each prefix product along the selected axis: every output element multiplies the corresponding input element by the preceding cumulative result. If axis is omitted, the input is flattened first.
Follow-up: How would you compute row-wise products for a 2-D array? Pass axis=1; axis=0 instead produces cumulative products down columns.
QWhat is the role of dtype?
dtype determines both the result array type and the accumulator type used during multiplication. If omitted, NumPy generally follows the input dtype, with a platform-integer rule for narrower integer inputs.
Follow-up: Why might dtype=np.float64 be useful? It can provide a wider numeric range and floating-point output, though floating-point overflow remains possible.
QHow does the out parameter work?
out is an optional ndarray that receives the computed result instead of requiring a newly allocated destination. It must have the expected shape and buffer length, and values may be cast to its type.
Follow-up: What should be checked before supplying out? Confirm its dimensions and capacity, and ensure its dtype can represent the intended result adequately.
Related
numpy.ndarray.cumprodnumpy.prodnumpy.cumsum