What is the numpy.intersect1d() function in Python?

Overview

NumPy is a popular library for working with arrays. NumPy’s intersect1d() function returns the intersection between two arrays. In other words, it returns the common elements for two given arrays.

Syntax

numpy.intersect1d(ar1, ar2, assume_unique=False, return_indices=False)

Parameters

This function accepts the following parameter values:

  • ar1 and ar2: These two required parameters represent the input arrays for which intersect1d() will return the intersection.

Note: intersect1d() accepts any array-like objects; this includes NumPy arrays and scalars.

Note: If any input array is not one-dimensional, the function will flatten them and convert them to a single dimensional array.

  • assume_unique: An optional parameter, passed as True if both input arrays are assumed to be unique and False otherwise. If both input arrays are unique, passing assume_unique as True can speed up calculation.
Note: If the input arrays are not unique and the user passes assume_unique as True, the function could return an incorrect result or an out-of-bound exception.
  • return_indices: An optional parameter, which determines if intersect1d() will return two extra arrays containing indices of the elements of the intersection array in the two input arrays.

Return value

  • The function always returns an array that includes the intersection elements found in both the input arrays; this is the intersection array from earlier.
  • The function optionally returns two additional arrays, which contain the indices of intersection elements in the input arrays. Each of these two optionally returned arrays represents one input array.
Note: The optional arrays are only returned when the return_indices input argument has been set to True.

Example

import numpy as np

# creating the input arrays

a = np.array([1,3,5,7,9])

b = np.array([2,4,6,8])



# finding the intersect of the two arrays

print(np.intersect1d(a, b))



# creating the input arrays

c = np.array([[1,2,3], [4,5,6]])

d = np.array([[1,2,3], [4,5,6]])



# finding the intersect of the two arrays

print(np.intersect1d(c, d))



# creating the input arrays

e = np.array([[1,2,3], [7,8,9]])

f = np.array([[1,2,3], [4,5,6]])



# finding the intersect of the two arrays

print(np.intersect1d(e, f, return_indices = True))

Explanation

  • Line 1: We import numpy as np.
  • Lines 3–4: We create two input arrays, a and b.
  • Line 7: We use intersect1d() to find the intersection of a and b, and print the results.
  • Lines 10–11: We create two input arrays, c and d. These are two 2D arrays.
  • Line 14: We use intersect1d() to find the intersection of c and d and print the results. The intersect1d() function returns a 1D array even though we input two 2D arrays.
  • Lines 17–18: We create two input arrays, e and f.
  • Line 21: We use intersect1d() to find the intersection of e and f, and print the results. The return_indices argument in intersect1d() has been set to True. As a result, intersect1d() returns two extra arrays, which contain indices of the intersection elements in the two input arrays. e and f both contain the intersection elements 1, 2, and 3 at indices 0, 1, and 2.

Free Resources

License: Creative Commons-Attribution-ShareAlike 4.0 (CC-BY-SA 4.0)