Python NumPy: How to Create Arrays

NumPy (Numerical Python) is the foundational library for scientific computing in Python. At its core lies the ndarray (n-dimensional array), a data structure that is significantly more efficient and powerful than standard Python lists for numerical operations. Creating these arrays correctly is the first step in any data science, machine learning, or scientific computing workflow.

This guide explores the essential methods for creating NumPy arrays, ranging from converting Python lists to generating data using built-in functions like zeros, ones, and arange.

Creating Arrays from Python Lists

The most direct way to create a NumPy array is by converting an existing Python list or tuple using the np.array() function.

1D and 2D Arrays

You can create a one-dimensional array from a simple list, or a two-dimensional (matrix) array from a list of lists.

import numpy as np

list_data = [1, 2, 3, 4, 5]
arr_1d = np.array(list_data)

print(f"1D Array: {arr_1d}")
print(f"Shape: {arr_1d.shape}")

matrix_data = [[1, 2, 3], [4, 5, 6]]
arr_2d = np.array(matrix_data)

print(f"\n2D Array:\n{arr_2d}")
print(f"Shape: {arr_2d.shape}")

Output:

1D Array: [1 2 3 4 5]
Shape: (5,)

2D Array:
[[1 2 3]
 [4 5 6]]
Shape: (2, 3)

Generating Arrays with Initial Placeholders

When you know the size of the array you need but not the specific data yet, use placeholder functions. This is more efficient than appending to a list in a loop.

Zeros and Ones

• np.zeros(shape): Creates an array filled with 0s.
• np.ones(shape): Creates an array filled with 1s.
• np.full(shape, value): Creates an array filled with a specific value.

import numpy as np

zeros = np.zeros((3, 3))
print(f"Zeros:\n{zeros}")

ones = np.ones((2, 4), dtype=int)
print(f"\nOnes (int):\n{ones}")

sevens = np.full((2, 2), 7)
print(f"\nFilled with 7:\n{sevens}")

Output:

Zeros:
[[0. 0. 0.]
 [0. 0. 0.]
 [0. 0. 0.]]

Ones (int):
[[1 1 1 1]
 [1 1 1 1]]

Filled with 7:
[[7 7]
 [7 7]]

Creating Arrays with Ranges and Intervals

NumPy provides powerful tools to generate sequences of numbers, similar to Python's built-in range().

np.arange vs np.linspace

• np.arange(start, stop, step): Like Python's range, but returns an array and supports floating-point steps.
• np.linspace(start, stop, num): Generates num evenly spaced values between start and stop (inclusive).

import numpy as np

range_arr = np.arange(0, 10, 2)
print(f"Arange: {range_arr}")

linear_arr = np.linspace(0, 1, 5)
print(f"Linspace: {linear_arr}")

Output:

Arange: [0 2 4 6 8]
Linspace: [0.   0.25 0.5  0.75 1.  ]

Creating Random Arrays

For simulation or machine learning initialization, you often need arrays populated with random numbers.

import numpy as np

rand_uniform = np.random.rand(2, 3) 
print(f"Random Uniform:\n{rand_uniform}")

rand_ints = np.random.randint(1, 100, size=(2, 2))
print(f"\nRandom Integers:\n{rand_ints}")

Output (values will vary):

Random Uniform:
[[0.10549263 0.50915867 0.7794045 ]
 [0.8860267  0.98328633 0.53472261]]

Random Integers:
[[64 81]
 [44 73]]

Common Pitfall: Array Dimensions and Shapes

A frequent error occurs when passing shapes to functions. Some functions expect dimensions as separate arguments, while others expect a tuple.

Example of Error

import numpy as np

try:
    err_array = np.zeros(2, 2) 
except TypeError as e:
    print(f"Error: {e}")

Output:

Error: Cannot interpret '2' as a data type

Solution

Pass dimensions as a single tuple (rows, cols).

import numpy as np

fixed_array = np.zeros((2, 2))
print(f"Correct Shape:\n{fixed_array}")

Output:

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

Conclusion

Creating NumPy arrays is the entry point to scientific computing in Python.

1. Use np.array() to convert existing lists.
2. Use np.zeros() or np.ones() to initialize arrays of a specific size.
3. Use np.arange() for step-based sequences and np.linspace() for point-based intervals.
4. Check your shapes: Ensure multi-dimensional sizes are passed as tuples (rows, cols).