NumPy ndarray.cumsum() Method: Cumulative Sum of Elements

NumPy ndarray.cumsum()

The numpy.ndarray.cumsum() method returns the cumulative sum of the elements in a NumPy array. It can compute the sum over the entire array or along a specified axis.

Syntax

</>

Copy

ndarray.cumsum(axis=None, dtype=None, out=None)

Parameters

Parameter	Type	Description
`axis`	None, int, optional	Specifies the axis along which the cumulative sum is computed. If `None`, the sum is computed over the entire array.
`dtype`	data-type, optional	Defines the type of the output array. If not specified, it uses the input array’s dtype.
`out`	ndarray, optional	Alternative output array for storing the result. Must have the same shape as the expected output.

Return Value

Returns a NumPy array of the same shape as the input, where each element is the cumulative sum of the preceding elements.

Examples

1. Computing the Cumulative Sum of a 1D Array

In this example, we compute the cumulative sum of a one-dimensional array.

</>

Copy

import numpy as np

# Creating a 1D NumPy array
arr = np.array([1, 2, 3, 4, 5])

# Computing the cumulative sum of the entire array
result = arr.cumsum()

# Printing the result
print(result)  # Output: [ 1  3  6 10 15]

Output:

[ 1  3  6 10 15]

Each element in the output represents the sum of all previous elements in the array up to that index.

2. Using the `axis` Parameter in ndarray.cumsum()

Here, we compute the cumulative sum along different axes in a 2D array.

</>

Copy

import numpy as np

# Creating a 2D NumPy array
arr = np.array([[1, 2, 3], 
                [4, 5, 6]])

# Cumulative sum along axis=0 (column-wise)
result_axis0 = arr.cumsum(axis=0)
print("Cumulative sum along axis 0 (column-wise):\n", result_axis0)

# Cumulative sum along axis=1 (row-wise)
result_axis1 = arr.cumsum(axis=1)
print("Cumulative sum along axis 1 (row-wise):\n", result_axis1)

Output:

Cumulative sum along axis 0 (column-wise):
[[ 1  2  3]
 [ 5  7  9]]

Cumulative sum along axis 1 (row-wise):
[[ 1  3  6]
 [ 4  9 15]]

For axis=0, the sum accumulates down each column, whereas for axis=1, the sum accumulates across each row.

3. Specifying Output Data Type Using `dtype`

We can specify the data type of the result using the dtype parameter.

</>

Copy

import numpy as np

# Creating an array with integer values
arr = np.array([1, 2, 3, 4], dtype=np.int32)

# Computing the cumulative sum and storing the result as float
result = arr.cumsum(dtype=np.float64)

print(result)  # Output: [ 1.  3.  6. 10.]

Output:

[ 1.  3.  6. 10.]

By specifying dtype=np.float64, the result is stored in floating-point format instead of integers.

4. Storing Output in a Specified Array Using `out`

The out parameter allows storing the result in an existing array.

</>

Copy

import numpy as np

# Creating an input array
arr = np.array([1, 2, 3, 4])

# Creating an output array of the same shape
out_arr = np.empty_like(arr)

# Computing cumulative sum and storing in out_arr
arr.cumsum(out=out_arr)

print(out_arr)  # Output: [ 1  3  6 10]

Output:

[ 1  3  6 10]

The cumulative sum is stored in out_arr instead of creating a new array.

TutorialKart

NumPy ndarray.cumsum() Method