w3resource

NumPy: Extract rows with unequal values from 10x3 matrix

Extract rows with unequal values.

Write a NumPy program to extract rows with unequal values (e.g. [1,1,2]) from 10x3 matrix.

Sample Solution:

Example-1:

Python Code:

# Importing the NumPy library
import numpy as np

# Generating a 2D NumPy array 'nums' filled with random integers from 0 to 3
nums = np.random.randint(0, 4, (6, 3))

# Displaying the original 2D array 'nums'
print("Original vector:")
print(nums)

# Checking for rows where consecutive elements are not equal within each row
# using np.logical_and.reduce and axis=1 to check row-wise
new_nums = np.logical_and.reduce(nums[:, 1:] == nums[:, :-1], axis=1)

# Filtering out rows with unequal consecutive values and storing in 'result'
result = nums[~new_nums]

# Displaying rows with unequal consecutive values
print("\nRows with unequal values:")
print(result)

Sample Output:

Original vector:
[[3 2 0]
 [2 3 1]
 [2 0 3]
 [3 3 1]
 [2 1 1]
 [3 0 2]]

Rows with unequal values:
[[3 2 0]
 [2 3 1]
 [2 0 3]
 [3 3 1]
 [2 1 1]
 [3 0 2]]

Explanation:

np.random.randint(0, 4, (6, 3)): Generates a 2D NumPy array nums with shape (6, 3) and random integers in the range of [0, 4).

new_nums = np.logical_and.reduce(nums[:,1:] == nums[:,:-1], axis=1)

In the above code -

  • nums[:, 1:]: Takes all rows and all columns starting from the second column (index 1) to the end of nums.
  • nums[:, :-1]: Takes all rows and all columns up to, but not including, the last column of nums.
  • nums[:, 1:] == nums[:, :-1]: Compares each pair of adjacent elements in each row. The result is a boolean array of the same shape as “nums” but with the last column removed.
  • np.logical_and.reduce(...): Performs a logical AND operation along the specified axis (axis=1, or columns) of the boolean array. This results in a 1D array with a True value for rows where all elements are equal and a False value otherwise.

Example-2:

Python Code:

import numpy as np
nums = np.array([(1,1,1),
              (1,1,1),
              (1,2,3)])
print("Original vector:")
print(nums)
new_nums = np.logical_and.reduce(nums[:,1:] == nums[:,:-1], axis=1)
result = nums[~new_nums]
print("\nRows with unequal values:")
print(result)

Sample Output:

Original vector:
[[1 1 1]
 [1 1 1]
 [1 2 3]]

Rows with unequal values:
[[1 2 3]]

Explanation:

result = nums[~new_nums]

In the above code -

  • ~new_nums: Negates the boolean array new_nums, inverting the True and False values.
  • nums[~new_nums]: Filters the rows in nums using the negated boolean array ~new_nums. This keeps the rows where not all elements are equal.

print(result): Finally print() function prints the resulting 2D NumPy array with rows from the original nums array where not all elements in a row are the same.

Python-Numpy Code Editor: