In numpy, “empty” is an ambiguous term because an empty array means an array with zero elements or an uninitialized array with arbitrary memory content. It can have various shapes (0,), (0, n), or (m, 0, n)—as long as its total number of elements is zero.

Creating an empty array

The optimal method for creating a one-dimensional empty array is to use the np.array() function.

For a multidimensional empty array, the np.empty() function is more efficient.

Using np.array()

You can directly initialize an array with zero elements like this:

import numpy as np

empty_array = np.array([])

print(empty_array)

print(empty_array.shape)

print(empty_array.size)

# Output:
# []
# (0,)
# 0

The output is a one-dimensional array with no values. Its shape is (0,). This approach is direct and readable.

Using np.empty()

For creating a 2D or multidimensional empty array, the better approach is to use the “np.empty()” function.

Let’s define rows and columns for a 2D array:

import numpy as np

arr2d = np.empty((0, 3)) # 0 rows and 3 columns

print(arr2d)

print(arr2d.size)

print(arr2d.shape)

# Output:
# []
# 0
# (0, 3)

It is important to note that the np.empty() method sometimes fills the rows with 0s, and sometimes it does not fill the array with zeros but with arbitrary uninitialized values based on the operating system.

import numpy as np

init_2d = np.empty((2, 3)) # 2 rows 3 columns

print(init_2d)

print(init_2d.size)

print(init_2d.shape)

# Output:

# [[0. 0. 0.]
#  [0. 0. 0.]]

# 6

# (2, 3)

For my macOS, it creates two rows and three columns of 0s. Your output might be different.

Let’s initialize a 3D empty array:

import numpy as np

arr3d = np.empty((2, 0, 4)) # 2 rows 0 columns 4 depth

print(arr3d)

print(arr3d.size)

print(arr3d.shape)

# Output:
# []
# 0
# (2, 0, 4)

Checking if a numpy array is empty

The most Pythonic way to check if an array is empty is to use the .size attribute. It returns the total number of elements in the array. If the size is 0, the array is empty; otherwise, it is not.

For 1D array

import numpy as np

empty_array = np.array([])

if empty_array.size == 0:
    print("Array is empty")
else:
    print("Array is not empty")

Output

Array is empty

For 2D array

For checking its emptiness, we can operate the .size attribute on it.

import numpy as np

# Empty 2D array
empty_2d_array = np.empty((0, 5))  # 0 rows, 5 columns
print(f"Empty 2D array:\n{empty_2d_array}")
if empty_2d_array.size == 0:
    print("The 2D array is empty (using .size)")

# Non-empty 2D array
non_empty_2d_array = np.array([[1, 2, 3], [4, 5, 6]])
print(f"\nNon-empty 2D array:\n{non_empty_2d_array}")
if non_empty_2d_array.size == 0:
    print("The 2D array is empty (using .size)")
else:
    print("The 2D array is NOT empty (using .size)")

Output

Empty 2D array:
[]
The 2D array is empty (using .size)

Non-empty 2D array:
[[1 2 3]
 [4 5 6]]
The 2D array is NOT empty (using .size)

It is extremely efficient and the least complex way.