NumPy empty_like(): Allocating Uninitialized Arrays from Prototype Geometries

The empty_like() function allocates memory instantly without initializing elements to zero or any other specific value. The resulting array contains arbitrary garbage values left in RAM from previous memory allocations. For zero-initialized cloning, see zeros_like(). For standalone uninitialized buffers, see empty().

Syntax and Parameters 🔝

numpy.empty_like(prototype, dtype=None, order='K', subok=True, shape=None)

Legacy vs Modernized Code Comparison 🔝

Review how legacy NumPy syntax and unformatted snippets compare against the modernized, structured, and self-verifying implementation:

Basic Prototype Cloning 🔝

Pass an existing array as a prototype to allocate a new uninitialized buffer sharing identical dimensions and data type:

import numpy as np

ar = np.array([4, 7, 9])
ar_empty = np.empty_like(ar)
print("Shape:", ar_empty.shape)
print("Dtype:", ar_empty.dtype)

assert ar_empty.shape == ar.shape
assert ar_empty.dtype == ar.dtype

Expected output (Note: Values will contain arbitrary memory contents)

Shape: (3,)
Dtype: int64

Overriding dtype 🔝

You can specify a custom dtype to change the element type of the newly allocated buffer while keeping the shape of the prototype:

ar = np.array([4, 7, 9])

# 1. Integer 64-bit override
ar_int64 = np.empty_like(ar, dtype=np.int64)
print("Int64 dtype:", ar_int64.dtype)

# 2. String dtype override
ar_str = np.empty_like(ar, dtype=str)
print("String dtype:", ar_str.dtype)

# 3. Float dtype override
ar_float = np.empty_like(ar, dtype=float)
print("Float dtype: ", ar_float.dtype)

assert ar_int64.dtype == np.int64
assert ar_float.dtype == np.float64

Expected output

Int64 dtype: int64
String dtype: 

Overriding Shape Geometries 🔝

The optional shape keyword argument allows you to allocate an uninitialized array with a brand new dimension while mirroring the prototype's dtype and memory layout:

ar = np.array([4, 7, 9])
ar_matrix = np.empty_like(ar, shape=(2, 3))

print("Overridden shape matrix:
", ar_matrix)
print("Shape:                   ", ar_matrix.shape)

assert ar_matrix.shape == (2, 3)
assert ar_matrix.dtype == ar.dtype

Expected output (Values represent uninitialized RAM bytes)

Overridden shape matrix:
 [[36598224        0        0]
 [       0        0        0]]
Shape:                    (2, 3)

Crucial Uninitialized Memory Warning 🔝

Performance Note & Risk: empty_like() and empty() do not zero-fill memory. They return whatever random bit pattern currently resides in that memory address in RAM. Always populate every element of the array immediately after allocation before reading from it to avoid undefined behavior or silent bugs.

Practice in Google Colab 🔝

Test prototype cloning, dtype overrides, shape adjustments, 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 array creation functions: empty(), zeros(), ones(), full(), eye(), and linspace().

« Array Creation empty() zeros() linspace() »

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:

← Subscribe to our YouTube Channel here