NumPy random.random_sample(): Uniform Random Floats in [0.0, 1.0)

import numpy as np

np.random.seed(42)
samples = np.random.random_sample(size=(2, 3))
print(samples)

The random_sample() function is fundamental for generating uniformly distributed random numbers, simulations, weight initializations in machine learning, and stochastic modeling. For random integers, see randint(). For normal distributions, see randn().

Syntax & Aliases 🔝

numpy.random.random_sample(size=None)
numpy.random.random(size=None)
numpy.random.ranf(size=None)
numpy.random.sample(size=None)

Legacy vs. Modernized Code Comparison 🔝

The table below compares legacy NumPy random snippets with the updated, structured implementation:

Generating 1D and Multi-Dimensional Arrays 🔝

Pass shape tuples to size to generate matrices of uniform random numbers:

import numpy as np

np.random.seed(100)

# 1D array of 5 floats
arr_1d = np.random.random_sample(size=5)
print("1D array:", arr_1d)

# 2D matrix (3 rows, 2 columns)
arr_2d = np.random.random_sample(size=(3, 2))
print("2D matrix:
", arr_2d)

assert arr_1d.shape == (5,)
assert arr_2d.shape == (3, 2)
assert np.all(arr_1d >= 0.0) and np.all(arr_1d < 1.0)

Expected output

1D array: [0.54340494 0.27836939 0.42451758 0.84477613 0.00471886]
2D matrix:
 [[0.12156912 0.67074904]
 [0.82585276 0.13670656]
 [0.57504433 0.89132198]]

Function Aliases: random, ranf, sample 🔝

NumPy provides identical functional aliases for random_sample(). np.random.random() is especially common in machine learning codebases:

import numpy as np

np.random.seed(42)

a = np.random.random(3)
b = np.random.ranf(3)
c = np.random.sample(3)

print("random():", a)
print("ranf():  ", b)
print("sample():", c)

assert np.array_equal(a, b) and np.array_equal(b, c)

Expected output

random(): [0.37454012 0.95071431 0.73199394]
ranf():   [0.37454012 0.95071431 0.73199394]
sample(): [0.37454012 0.95071431 0.73199394]

Scaling to Custom Ranges [Low, High) 🔝

To generate uniform random floats across an arbitrary range [low, high), multiply the standard [0.0, 1.0) samples by the range width and add the lower bound:

low, high = 5.0, 15.0
scaled = (high - low) * np.random.random_sample(4) + low
print("Scaled floats in [5, 15):", scaled)

assert np.all(scaled >= low) and np.all(scaled < high)

Expected output

Scaled floats in [5, 15): [ 8.74540119 14.50714306 12.31993936  6.15858262]

Modern Generator: default_rng().random() 🔝

In modern NumPy (v1.17+), the recommended approach uses the Philox/PCG64 bit generator via default_rng() rather than legacy global state functions:

rng = np.random.default_rng(seed=42)
modern_samples = rng.random(size=(2, 2))
print("Modern Generator samples:
", modern_samples)

assert modern_samples.shape == (2, 2)

Expected output

Modern Generator samples:
 [[0.77395605 0.43887844]
 [0.85859792 0.69736803]]

Common Pitfalls & Best Practices 🔝

• Exclusive Upper Bound: Remember that values generated are strictly less than 1.0 (interval is [0.0, 1.0)).
• Shape Tuple Syntax: When requesting multi-dimensional arrays, pass dimensions as a tuple size=(rows, cols) rather than separate positional arguments.

Practice in Google Colab 🔝

Test random_sample() arrays, function aliases, custom range scaling, modern generators, and run verification test suites in our interactive Google Colab companion notebook:

Interactive Practice: Colab & GitHub

Launch the notebook directly in Google Colab to run live code cells or inspect the source repository on GitHub:

View on GitHub

Continue exploring related NumPy random functions: randn(), randint(), and NumPy Overview.

Subhendu Mohapatra

Author & Instructor at plus2net

I write and maintain practical tutorials on Python, PHP, SQL, JavaScript, HTML, jQuery, and web development at plus2net. The tutorials focus on clear explanations, working examples, and code that readers can test and adapt while learning.

About the author   Watch on YouTube

Connect: