np.full() for custom-filled arrays in NumPy

Introduction

We use np.full() to quickly create arrays filled with the same value. This helps when you want a starting point or a fixed value repeated many times.

When you want an array filled with zeros, ones, or any specific number for calculations.

When you need a placeholder array before filling it with real data.

When testing algorithms that require input arrays with constant values.

When initializing weights or biases in simple machine learning models.

When creating masks or filters with a fixed value.

Syntax

NumPy

import numpy as np

array = np.full(shape, fill_value, dtype=None, order='C')

shape is the size of the array you want, like (3, 4) for 3 rows and 4 columns.

fill_value is the number you want to fill the array with.

Examples

This creates a 2 rows by 3 columns array where every element is 7.

NumPy

import numpy as np

# Create a 2x3 array filled with 7
array = np.full((2, 3), 7)
print(array)

Shows what happens if the shape is empty: the result is an empty array.

NumPy

import numpy as np

# Create an empty array with shape (0,) filled with 5
array = np.full((0,), 5)
print(array)

Creates a single-element array with the value 10.

NumPy

import numpy as np

# Create a 1-element array filled with 10
array = np.full((1,), 10)
print(array)

Creates a 3x3 array filled with 0.5 as floats.

NumPy

import numpy as np

# Create a 3x3 array filled with 0.5 and specify float type
array = np.full((3, 3), 0.5, dtype=float)
print(array)

Sample Program

This program creates a 4 by 5 array filled with 9. Then it changes one element to 0 to show how you can modify the array after creating it.

NumPy

import numpy as np

# Create a 4x5 array filled with 9
array_before = np.full((4, 5), 9)
print("Array before change:")
print(array_before)

# Change the value at position (2,3) to 0
array_before[2, 3] = 0
print("\nArray after changing one element to 0:")
print(array_before)

Important Notes

Time complexity is O(n) where n is the total number of elements, because it fills each element.

Space complexity is O(n) because it creates a new array of the given size.

A common mistake is to forget that shape must be a tuple, even for 1D arrays (use (5,) not 5).

Use np.full() when you want a fixed value repeated. Use np.zeros() or np.ones() if you only need zeros or ones.

Summary

np.full() creates arrays filled with the same value.

It needs the shape and the fill value as inputs.

Useful for initializing arrays with constant values for calculations or placeholders.

Practice

1. What does the np.full() function do in NumPy?

easy

A. Creates an empty array with no values

B. Creates an array with random values

C. Creates an array with increasing numbers

D. Creates an array filled with the same specified value

Solution

1. Step 1: Understand the purpose of np.full()

   The function np.full() is designed to create arrays where every element has the same value.

2. Step 2: Compare with other array creation functions

   Unlike functions that create random or sequential arrays, np.full() fills the array with a constant value.

3. Final Answer:

   Creates an array filled with the same specified value -> Option D

4. Quick Check:

   np.full() = constant value array [OK]

Hint: Remember: full means fill with one value [OK]

Common Mistakes:

• Confusing np.full() with np.random functions
• Thinking it creates empty or zero arrays
• Assuming it creates sequences of numbers

2. Which of the following is the correct syntax to create a 3x3 NumPy array filled with the value 7 using np.full()?

easy

A. np.full((3, 3), 7)

B. np.full(3, 3, 7)

C. np.full([3, 3], value=7)

D. np.full(7, shape=(3,3))

Solution

1. Step 1: Identify the correct parameter order for np.full()

   np.full() takes the shape as the first argument (tuple), then the fill value as the second argument.

2. Step 2: Check each option's syntax

   np.full((3, 3), 7) uses np.full((3, 3), 7) which matches the correct order and types. Others misuse argument order or syntax.

3. Final Answer:

   np.full((3, 3), 7) -> Option A

4. Quick Check:

   Shape tuple first, fill value second [OK]

Hint: Shape is first tuple, value is second argument [OK]

Common Mistakes:

• Passing shape and value as separate arguments without tuple
• Using keyword arguments incorrectly
• Swapping the order of shape and fill value

3. What is the output of the following code?

import numpy as np
arr = np.full((2, 2), 5)
print(arr)

medium

A. [[5 5]
[5 5]]

B. [[2 2]
[2 2]]

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

D. [[5 5 5]
[5 5 5]]

Solution

1. Step 1: Understand np.full() parameters

   The shape is (2, 2), so the array will have 2 rows and 2 columns. The fill value is 5.

2. Step 2: Predict the printed array

   The array will be a 2x2 matrix with all elements equal to 5, printed as [[5 5]\n [5 5]].

3. Final Answer:

   [[5 5] [5 5]] -> Option A

4. Quick Check:

   Shape (2,2) with 5 fills = [[5 5]\n [5 5]] [OK]

Hint: Shape defines size, fill value repeats [OK]

Common Mistakes:

• Confusing shape dimensions
• Expecting zeros instead of fill value
• Misreading the printed array shape

4. Identify the error in this code snippet:

import numpy as np
arr = np.full(3, 4, 2)
print(arr)

medium

A. Fill value must be a string

B. Shape argument must be a tuple

C. The third positional argument `2` is an invalid dtype specifier

D. np.full() cannot create 1D arrays

Solution

1. Step 1: Check np.full() argument count and order

   np.full(shape, fill_value, dtype=None, ...). The call np.full(3, 4, 2) sets shape=3, fill_value=4, dtype=2, but 2 is not a valid dtype (raises ValueError: data type "2" not understood).

2. Step 2: Confirm correct usage

   Correct syntax is np.full(3, 4) for 1D array of 4s, or specify valid dtype like dtype=float. Extra or invalid dtype causes error.

3. Final Answer:

   The third positional argument `2` is an invalid dtype specifier -> Option C

4. Quick Check:

   np.full(3,4,2) -> ValueError: data type "2" [OK]

Hint: Optional dtype must be valid (e.g., float, int) [OK]

Common Mistakes:

• Passing multiple positional arguments instead of tuple shape
• Using wrong data types for shape
• Assuming np.full() needs three arguments

5. You want to create a 4x4 NumPy array filled with the value 9, but only the diagonal elements should be 0. Which code correctly achieves this using np.full() and other NumPy functions?

hard

A. arr = np.zeros((4,4)); arr[:] = 9; print(arr)

B. arr = np.full((4,4), 9); np.fill_diagonal(arr, 0); print(arr)

C. arr = np.full((4,4), 0); np.fill_diagonal(arr, 9); print(arr)

D. arr = np.full(4, 9); np.fill_diagonal(arr, 0); print(arr)

Solution

1. Step 1: Create a 4x4 array filled with 9

   Using np.full((4,4), 9) creates a 4x4 array where every element is 9.

2. Step 2: Replace diagonal elements with 0

   np.fill_diagonal(arr, 0) sets the diagonal elements of the array to 0.

3. Step 3: Verify the output

   Printing arr shows a 4x4 array with 9 everywhere except 0 on the diagonal.

4. Final Answer:

   arr = np.full((4,4), 9); np.fill_diagonal(arr, 0); print(arr) -> Option B

5. Quick Check:

   Fill then fill_diagonal for custom values [OK]

Hint: Fill full array first, then change diagonal [OK]

Common Mistakes:

• Filling zeros first then trying to fill diagonal with 9
• Using wrong shape or 1D array for 2D problem
• Misusing np.fill_diagonal on 1D arrays