np.random.random() returns random floats sampled uniformly from the half-open interval [0.0, 1.0), functionally identical to np.random.rand() but taking a single shape argument (a tuple or integer) instead of separate dimension arguments.

Syntax

np.random.random(size=None)

Deep Dive Course

random() and rand() draw from exactly the same underlying uniform distribution and the same random state, differing only in their calling convention: random() takes size as one argument, following NumPy's usual shape-as-tuple/int pattern, while rand() takes each dimension as a separate positional argument. random() exists partly for consistency with Python's built-in random.random() function, which also returns a single uniform value in [0, 1).

1Understanding np.random.random()

💡

random() and rand() are functionally interchangeable in terms of distribution — pick based on which calling convention you prefer, size as a tuple with random(), or dimensions as separate arguments with rand().

editor.html

import numpy as np

np.random.seed(0)
print(np.random.random(3))

localhost:3000

2Practical Example

Here is a real-world application of np.random.random() showing how it is used in production NumPy code.

editor.html

import numpy as np

np.random.seed(0)
matrix = np.random.random((2, 2))
print(matrix)

localhost:3000

3Best Practices

Follow these guidelines when working with np.random.random():

1. Use np.random.random(size) when a shape tuple is more convenient or consistent with surrounding code that uses other shape-taking functions

2. Avoid mixing rand() and random() inconsistently within the same codebase, since they behave identically but read differently

3. Seed explicitly whenever a specific sequence needs to be reproducible for tests or demonstrations

⚠️

Tip: random() and rand() are functionally interchangeable in terms of distribution — pick based on which calling convention you prefer, size as a tuple with random(), or dimensions as separate arguments with rand().

editor.html

import numpy as np

np.random.seed(0)
print(np.random.random(3))

localhost:3000

Examples

Example 01Basic Usage

import numpy as np

np.random.seed(0)
print(np.random.random(3))

Example 02Advanced Example

import numpy as np

np.random.seed(0)
matrix = np.random.random((2, 2))
print(matrix)

Best Practices

Use np.random.random(size) when a shape tuple is more convenient or consistent with surrounding code that uses other shape-taking functions
Avoid mixing rand() and random() inconsistently within the same codebase, since they behave identically but read differently
Seed explicitly whenever a specific sequence needs to be reproducible for tests or demonstrations

Interview Question

What's the actual difference between np.random.random() and np.random.rand(), given they sample from the same distribution?▼

Hint: Think about how each function accepts its shape argument.

Both functions draw uniformly from [0, 1) using the exact same underlying random state and algorithm — there's no statistical difference between them. The only difference is the calling convention: random() accepts a single size argument, which can be an integer or a shape tuple, matching the convention used by most other NumPy array-creation functions, while rand() instead accepts each dimension as a separate positional argument, an older, special-cased style shared with a few of NumPy's other legacy random functions.

Exercises

MediumPractice using np.random.random() in a real scenario.

View Solution

import numpy as np

np.random.seed(0)
print(np.random.random(3))

Frequently Asked Questions

What's the actual difference between np.random.random() and np.random.rand(), given they sample from the same distribution?▼

Related Functions

np-random-randnp-random-seednp-random-uniform